article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
several recent observations indicate that even the low h i column density @xmath0 absorption lines may not be primordial as were previously thought .
high resolution observations by cowie ( 1995 ) , tytler ( 1995 ) and womble ( 1996 ) show that there are metal lines ( c iv ) associated with @xmath0 absorption lines with as small h i column density as @xmath11 @xmath7 at @xmath5 .
the fraction of the @xmath0 lines with @xmath12 @xmath7 that are associated with metal lines is approximately @xmath13 at @xmath5 .
the median abundance of lines with @xmath14 @xmath7 at @xmath5 , is found to be [ c / h ] @xmath15 .
recently songaila and cowie ( 1996 ) have shown that the fraction of @xmath0 lines that are associated with metal lines , hereafter denoted as @xmath16 , could be higher , and they have also detected the other c ii , s iv , n v absorption lines in hydrogen clouds with @xmath17 @xmath7 . in view of the dependence of observational sensitivity for detecting metal lines , the fraction @xmath16 represents only a lower limit for the fraction of @xmath0 lines associated with ` metals ' .
another set of observations shows that at low redshifts a large fraction of @xmath0 absorption lines are associated with galaxies ( maloney 1992 ; morris 1993 ) .
lanzetta ( 1995b , lbtw ) claimed that @xmath18 of the @xmath0 lines at @xmath4 arise in galactic halos .
the observation of barcons ( 1995 ) in which they imaged a galaxy corresponding to an absorption line , is an evidence of such associations .
it is , therefore , reasonable to assume that at least a fraction of the @xmath0 absorption lines arise in galactic halos , and this fraction is hereafter denoted as @xmath19 . in the hierarchical model of structure formation
, there is a continuum of halos of various masses and velocity dispersions . in this paper
, we consider the hypothesis that the @xmath0 lines are produced in two populations of halos in this continuum in ` minihalos ' of small velocity dispersions which have not cooled and formed stars , and in large galactic halos with stars and metals . based on this two - population model for qso absorbers , we attempt to explain @xmath16 in terms of @xmath19 by the relation @xmath20
. there are circumstantial evidences for such a two population scenario . in the column density distribution of @xmath0 lines , there is a break around @xmath21 @xmath7 ( see , e.g. , cristiani 1995 ) .
petitjean ( 1993 ) have interpreted this break as being due to existence of two populations of @xmath0 lines
the lines with higher h i column density being associated with strong metal lines and with galactic halos .
recently , fernndez - soto et al . ( 1996 ) analyzed the two - point correlation of very weak c iv absorption lines associated with high - redshift ( @xmath22 ) @xmath0 absorption systems .
they found that high - redshift @xmath0 absorption systems traced by c iv lines are clustered in redshift , as strongly as that expected for galaxies .
this suggests that many @xmath0 absorbers even at high redshifts may arise in galaxies .
our primary goal in this paper is therefore to determine whether these recent observations can be understood within the framework of a two - population model for @xmath0 absorption systems , including their redshift evolution in the context of hierarchical structure formation . in 2 , we present a model of @xmath0-absorbing gas confined in both mini and galactic halos and set the prescription for determining the number of absorption lines provided by various masses of halos . on the basis of the press - schechter mass function with various power spectra and cosmological parameters ,
we then discuss the fraction of absorption lines in minihalos and in galactic halos by comparing with the recent keck observations in 3 . the simplified approach for modeling chemical evolution of halos
is also presented to predict the evolution of the average metallicity associated with @xmath0 lines in redshift .
the discussion and conclusions are finally drawn in 4 .
we define minihalos as halos in which baryonic gas is entirely photoionized by the external uv background and the gas is stably confined in the gravitational field of the dark matter ( rees 1986 ; ikeuchi 1986 ) .
in these minihalos , the gas does not go through extensive star formation because of suppression of radiative cooling by photoionization ( efstathiou 1992 ) .
below , we will characterise the halos by their circular velocities @xmath9 , where the mass density of halos is dominated by extended dark matter with @xmath23 distribution .
minihalos are therefore bounded by two scales of circular velocity , @xmath24 .
the lower limit is set by the sound velocity of the photoionized intergalactic medium ( igm ) , which is @xmath25 km s@xmath2 .
halos with @xmath9 less than this value will not have baryonic gas infall from the igm .
this is basically the jeans mass , below which the perturbation in baryonic gas is suppressed .
we take a characteristic value of @xmath26 km s@xmath2 for the photoionized igm with @xmath27 k. the upper limit @xmath28 is admittedly less certain .
recent numerical simulations show that photoionization severely inhibits the formation of galaxies with @xmath29 km s@xmath2 , and substantially decreases the mass of cooled baryons in halos with @xmath30 km s@xmath2 ( thoul and weinberg 1996 ) .
they also find that above @xmath31 km s@xmath2 photoionization of the igm does not have any significant effect .
for the reasons described in the next subsection , we will use a canonical value of @xmath32 km s@xmath2 and examine the effect of changing @xmath28 over its possible range .
in fact , we later ask the question whether or not one can _ determine _ the value of @xmath28 from the observations . in brief , then , we assume that halos with @xmath33 km s@xmath2 do not undergo vigorous star formation , and call them minihalos in this paper .
we will call the halos with circular velocity larger than @xmath28 , the galactic halos .
our results are not sensitive to the upper limit of the circular velocity for galactic halos , and we will use a value of @xmath34 km s@xmath2 as an upper limit .
below , we will associate these halos with @xmath0 lines which have corresponding metal lines .
it is true that the division of halos into the above categories of mini and galactic halos is rather ad hoc .
however , it is most likely to be close to the real picture for the above mentioned reasons . for the gas associated in galactic halos , the virial temperature @xmath35 for @xmath36 km
s@xmath2 is above @xmath37 k [ @xmath38 k ] , where the mean molecular weight is taken as 0.59 for a totally ionized gas . in this temperature range ,
any local disturbances in the hot gas may be thermally unstable , because the radiative cooling rate increases with decreasing @xmath35 .
we therefore assume that the instability promotes the development of a two - phase structure in the gas ( fall & rees 1985 ) , so that it is the cool gas clouds , being pressure - confined by the surrounding hot gas , that produce @xmath0 absorptions , as proposed by mo ( 1994 ) . in this work
, we will be concerned with @xmath0 lines with low hi column density as @xmath39 @xmath7 , in which the shielding against the uv background radiation is ineffective .
the internal density of clouds in the outer parts of halos that give rise to such lines is low enough to be photoionized by the uv background radiation .
we thus assume the gas in clouds to be in photoionization equilibrium , where the gaseous temperature is expected to be order of @xmath40 k , and therefore consistent with the observed doppler parameters of @xmath0 lines ( see , e.g. , maloney 1992 ; rauch et al .
1996 ) .
note that , even the gas in disks , if disks are formed , will be ionized by the uv radiation in the outer parts , beyond the observed sharp edge at @xmath41 @xmath7 ( maloney 1992 ) .
the galaxies at low redshift imaged by barcons ( 1995 ) are inferred to give rise to absorption lines at an impact parameters of @xmath42 kpc , where @xmath43 km s@xmath2 mpc@xmath2 .
the gas which is responsible for such absorption lines is certainly ionized by the uv background radiation ( see also , morris 1993 ; salpeter and hoffmann 1995 ) .
we assume that the density profile of both mini and galactic halos is represented by that of a softened isothermal sphere , @xmath44 where @xmath45 is the central density and @xmath46 is the core radius .
the baryonic gas amounts to a fraction @xmath47 of the total mass of the halo .
then , the circular velocity @xmath48 at @xmath49 can be written as , @xmath50 where @xmath51 .
this functional form of @xmath48 assures a nearly flat rotation curve at large radii @xmath52 . in what follows
, the circular velocity is characterised by its value at the virial radius of the halo , @xmath53 ( see below for an explanation of @xmath53 ) . then , substituting eq.(2 ) into eq.(1 ) and introducing @xmath54 , the density profile of a halo that includes both baryonic gas and dark matter is re - written as , @xmath55 where @xmath9 is the circular velocity at @xmath56 .
we model the gas in minihalos ( @xmath57 ) to be in photoionized equilibrium with the external uv background radiation .
we write the intensity of the uv background radiation at the lyman limit as @xmath58 erg @xmath7 s@xmath2 sr@xmath2 hz@xmath2 . using the ionization and recombination rates for a fiducial uv spectrum of @xmath59 ( black 1981 ) , the h i number density @xmath60 in minihalos
is written as , @xmath61 where @xmath62 for hydrogen and @xmath63 is the recombination rate coefficient ( black 1981 ) .
the column density of h i at an impact parameter @xmath64 is then estimated as @xmath65 ^ 2 } \quad { \rm cm}^{-2 } \quad { \rm for\ minihalos } \
% % & ( 5)\end{aligned}\ ] ] for galactic halos ( @xmath66 ) , we assume a cloud confined by ambient hot gas to be a uniform sphere with proton number density @xmath67 and temperature @xmath68 . the pressure balance with hot gas with proton number density @xmath69 ( @xmath70 ) and temperature @xmath35 yields @xmath71 . under the assumption that the clouds are photoionized by the external uv background and in ionization equilibrium ,
the h i number density @xmath60 is then written as @xmath72 the cloud radius @xmath73 for the mass of each cloud @xmath74 is @xmath75 each cloud produces the h i column density as @xmath76 . in order to obtain the h i column density observed at an impact parameter @xmath64 from the galactic center , it is required to know the spatial distribution of clouds in halos _ a priori_. without going into details for modeling this , we assume that the clouds have roughly unit covering factor in halos when projected onto the sky , and that each sightline encounters only one cloud in average , which is not in disagreement with the observational results ( morris et al .
1993 ; lbtw ) . then
, as we argue in the appendix , the average h i column density observed at an impact parameter @xmath64 is well approximated by @xmath77 .
we thus obtain from eq.(6 ) and ( 7 ) @xmath78^{5/3 } } \quad { \rm cm}^{-2 } \quad { \rm for\ galactic\ halos } \ > , % % & ( 8)\end{aligned}\ ] ] where we assume the virial temperature for @xmath35 of the hot gas .
we note that both expressions for @xmath79 given in eq.(5 ) and ( 8) are approximately connected continuously at the canonical value of @xmath80 km s@xmath2 for @xmath28 , corresponding to the virial temperature of @xmath81 k ; this is the temperature of hot gas below which the picture of a two - phase structure in the gas becomes invalid .
we therefore use eq.(5 ) for @xmath82 km s@xmath2 and eq.(8 ) for @xmath83 km s@xmath2 in the following . according to several determinations of the intensity of the uv background radiation ( e.g. madau 1992 ) ,
@xmath84 is nearly constant at high redshifts @xmath85 , whereas at low redshifts @xmath86 it rapidly decreases with decreasing redshift .
we adopt the following functional form for @xmath84 , @xmath87 where the index @xmath88 quantifies the decreasing rate of @xmath84 for @xmath89 . to estimate the value of @xmath90 in eq.(3 ) , we use the theory of the spherical collapse of cosmological density perturbations , that yields @xmath91 for the einstein - de sitter ( eds ) universe ( e.g. peebles 1980 ; padmanabhan 1993 ) .
this allows us to derive @xmath92 using the observed values of @xmath9 and @xmath46 of the present - day dark matter halos at @xmath93 . for this purpose , it is convenient to use dark - matter dominated dwarf spirals ( moore 1994 ) , where the dynamical effect of a luminous component on the observed rotation curve is sufficiently small , so that one can accurately determine the structural parameters of dark halos . adopting the values of @xmath9 and @xmath46 from moore ( 1994 ) for dwarf spirals of ddo154 , ddo170 , ddo105 and ngc3109 , we obtain @xmath94 for the eds universe with @xmath95 . in
what follows , we assume @xmath96 , since the following results in the concerned low - column density range of @xmath97 @xmath7 are quite insensitive to the values of @xmath92 as long as @xmath98 .
figure 1(a ) shows the impact parameter @xmath64 _ vs. _ the total mass of halos @xmath99 at @xmath100 , where we take @xmath101 k for the gas in minihalos ( eq.5 ) , @xmath102 k and @xmath103 for the clouds in galactic halos ( eq.8 ) , with the common parameters of @xmath104 , @xmath105 , @xmath106 , and @xmath107 .
it is seen that lines of constant @xmath79 have the slightly different slope at @xmath108 ( or @xmath1 km s@xmath2 ) from that at larger @xmath99 or @xmath9 .
this is because @xmath109 in eq.(5 ) while @xmath110 in eq.(8 ) for @xmath111 .
here we note that the discontinuities at the boundary of both regimes ( at @xmath80 km s@xmath2 ) are quite small in @xmath64 so that these will not affect the later results .
we also find that h i column densities as low as @xmath112 @xmath7 can be provided by either the low - mass minihalos with the impact parameter @xmath64 below @xmath113 kpc or by the high - mass galactic halos with larger @xmath64 .
this is more clearly presented in fig.1(b ) that shows @xmath79 _ vs. _
@xmath64 for a given total mass @xmath99 .
dotted line denote the observed size of gaseous envelopes @xmath114 @xmath115 kpc in luminous galaxies which give rise to @xmath0 absorptions at @xmath4 ( lbtw ) . in our model of absorbers , this is achieved for a mass range of @xmath116 @xmath117 , which corresponds to galaxies with luminosity @xmath118 @xmath119 for the baryon fraction @xmath47 of 0.05 and the mass - to - light ratio of the order of 5 .
thus , the model yields the right order of the size of extended @xmath0-absorbing gaseous envelopes in galaxies .
bold solid line in fig.1(b ) corresponds to the observed correlation between the h i column density and impact parameter by lbtw , assuming doppler parameters of absorption lines in the range of @xmath120 km s@xmath2 .
the present model also implies the correlation and agrees approximately with lbtw , where the slight discrepancy in the slope may be due to the effects of sampling over different galactic masses and/or impact parameters .
recently le brun ( 1996 ) and bowen ( 1996 ) claimed that the equivalent width of @xmath0 lines do not correlate with the impact parameter .
this has been argued to weaken the possibility of @xmath0 lines being directly associated with galaxies .
the other possibilities discussed in the literature include the scenario of filaments between luminous galaxies , and @xmath0 lines from outflows from galaxies .
the issuse may be settled by assembling the statistically meaningful number of @xmath0 absorber - galaxy pairs and by comparing with the current type model that presents the explicit dependence of an impact parameter on the masses of galaxies and redshifts ( fig.1c ) . with the knowledge of the h i column density that is produced at an impact parameter @xmath121
, we can calculate the number of @xmath0 lines with a column density larger than @xmath122 , if the number density of halos with circular velocity @xmath9 is known .
we use the press - schechter formalism for this purpose , and cdm models of structure formation ( similar to the work of mo et al .
1993 ) . for the eds universe ,
the circular velocity @xmath9 and mass @xmath99 of a cdm halo are related as @xmath123 where @xmath124 is the present - day mean density of the universe and @xmath125 is the comoving radius of the halo .
the comoving number density of halos per unit circular velocity @xmath9 is given by ( mo et al .
1993 ) @xmath126 here @xmath127 and @xmath128 are the rms and threshold linear overdensities in a spherical region of radius @xmath125 , respectively , extrapolated to the present epoch . with a top - hat filtering , the functional form of @xmath127 for the cdm power spectrum of density perturbations
is taken from bardeen et al .
( 1986 ) , and its value is normalized at a scale of @xmath129 mpc with the bias parameter @xmath130 . the number of absorption systems per unit redshift with an h i column density larger than @xmath122 , that is produced in halos with circular velocity @xmath131 , is then given as ( sargent et al .
1980 ; mo et al . 1993 ) , @xmath132 for the eds universe .
we define @xmath133 for minihalos and @xmath134 for galactic halos . here
@xmath135 denotes the fraction of halos that give rise to @xmath0 absorptions , in view of the fact that some fraction of halos , specifically early - type galaxies such as e , s0 , de and ds0 in the scale of galactic halos , are lacking of interstellar gas .
since these galaxies experienced the burst of star formation and the subsequent gas loss via galactic wind for only a short period of @xmath136 gyr ( see e.g. yoshii & arimoto 1987 ) , the probability of a sightline to pass the protogalactic , gas - rich stage of these galaxies may be small . to evaluate @xmath135
, we adopt the nominal fraction 69% of late - type galaxies among all types derived by postman and geller ( 1984 ) and assume it to be constant with @xmath137 for simplicity .
thus , @xmath138 then the above equation can be used to find the fraction of @xmath0 lines with @xmath139 that are associated with galactic halos : @xmath140 besides the case of the ` standard ' eds universe , we will also calculate the number of @xmath0 lines @xmath141 and their galactic fraction @xmath19 for a world model with a cosmological constant .
the basic cosmological parameters that define the evolution of the universe are the density parameter @xmath142 , where @xmath143 is the critical density to close the universe , and the cosmological constant @xmath144 .
we investigate a low - density ( @xmath145 ) , flat ( @xmath146 ) universe , since such a universe deserves special attention to solve several paradoxical results of cosmological observations ( ostriker & steinhardt 1995 ) .
we use the corresponding set of equations ( 10)-(12 ) derived from the spherical collapse theory of density perturbations in a non - zero @xmath147 universe ( see peebles 1980 and suto 1993 for details ) . the model universes which we adopt for the following analysis are @xmath148 and @xmath149 .
both universes yield the cosmological age as @xmath150 gyr , so that any differences in the final results , which may arise from the different cosmological age , are eliminated .
we first calculate the number of @xmath0 absorption lines based on the model described in the last section , and investigate whether or not the model results can be reconciled with observations .
figure 2(a ) shows the redshift evolution of the number of @xmath0 absorption lines with an h i column density larger than @xmath151 @xmath7 .
the absorption lines arise from both minihalos and galactic halos , and the bounds on circular velocities are taken as @xmath26 km s@xmath2 and @xmath34 km s@xmath2 .
we take a standard set of parameters for photoionized gas ( @xmath101 k for minihalos , @xmath102 k and @xmath103 for galactic halos , with @xmath104 and @xmath105 ) .
solid lines correspond to a ` standard ' eds universe ( @xmath152 , @xmath153 ) , while dotted lines to a low - density ( @xmath154 ) , flat ( @xmath146 ) universe .
it is found , as was already argued by mo et al .
( 1993 ) , that there are two stages in the evolution of the number density .
first , the rapid formation of minihalos inherent in the cdm models gives rise to the abundant of @xmath0 lines at @xmath155 , and second , at lower redshifts of @xmath156 , the number of @xmath0 lines decreases with time as the hierarchical mergings proceed small - sized , less - massive objects disappear by being incorporated into larger , more massive objects .
the rapid decrease of @xmath141 with decreasing @xmath137 is now an observationally established fact , first recognized by peterson ( 1978 ) . for the cases of large @xmath130 and non - zero @xmath147 ,
the maximum value of @xmath141 is reduced and the epoch at which it is realized is delayed .
this is because the growth of density perturbations and subsequent mergings are suppressed due to the bias in galaxy formation and the rapid expansion of the universe .
however , the effects of these different assumptions on model parameters appear to be almost indistinguishable at the second stage of the number evolution of @xmath0 lines .
this indicates that probing world models by @xmath141 of absorption systems in the observed redshift region of @xmath157 may not be useful in view of some model uncertainties .
we note here that numerical simulations with a variety of world models have shown almost similar h i column density distribution for @xmath0 lines ( miralda - escud 1995 ; hernquist 1995 ) .
also plotted in the figure as filled squares are the observed numbers of @xmath0 lines at @xmath5 ( petijean et al .
1993 ) , @xmath158 ( lu et al .
1991 ) , and @xmath159 ( bahcall et al . 1993 ) .
it is clear that the models with @xmath160 at @xmath161 are reasonably in agreement with the observed values in the corresponding range of redshifts , whereas the _ hubble space telescope _ ( hst )
result of the unexpectedly large number of low - redshift @xmath0 lines ( bahcall et al . 1993 ) , compared to the extrapolated value from high @xmath137 , can be explained if the uv flux is decreasing with time ( mo et al . 1993 ) .
we find that @xmath162^\alpha$ ] with @xmath163 fits well .
it is interesting to note that with such a redshift evolution , the current uv flux is @xmath164 , which is consistent with the observations of maloney ( 1992 ) and kulkarni and fall ( 1993 ) .
the dependence of @xmath141 on the column density @xmath79 is shown in fig.2(b ) , where filled squares are taken from rauch et al .
( 1992 ) for @xmath165 .
the observed power - law decline of the @xmath0 lines , @xmath166 , is well reproduced by the model on the basis of the isothermal density profile of minihalos ( rees 1986 ) , together with the current model of pressure - confined clouds in galactic halos .
we remark that this is the case as long as the effect of self - shielding against the external uv radiation is neglected .
the effect , leaving a cooled neutral core in the inner part of a sphere ( e.g. chiba & nath 1994 ) , is realized at the high column - density ends of @xmath167 @xmath7 .
this effect is thus unimportant in the present study of low column - density @xmath0 lines of @xmath168 @xmath7 that occur at the large impact parameters of sightlines ( fig.1 ) .
we now investigate what fraction of @xmath0 lines can be associated with galactic halos where production of metals is possible the fraction @xmath19 which may be compared to that of @xmath0 lines associated with metal lines , @xmath16 .
figure 3(a ) shows the evolution of the galactic fraction @xmath19 of @xmath0 lines with @xmath169 @xmath7 , where we assume @xmath170 km s@xmath2 as the boundary value of @xmath9 between minihalos and galactic halos .
it is found that initially at high redshifts ( @xmath171 ) , most of @xmath0 lines are provided by minihalos that collapsed early for the cdm - type power spectrum of density perturbations .
this leads to the small values of @xmath19 less than 0.2 .
then , at @xmath172 , the fraction of lines in galactic halos is rapidly increasing with time as a result of the formation of massive halos via hierarchical merging .
finally at @xmath156 , the value of @xmath19 slowly converges to the current fraction of @xmath173 .
the figure also indicates that the differences arising from the different bias parameters @xmath130 and cosmological parameters @xmath174 , which are clearly visible at high redshifts , become less significant at lower redshifts where observations of absorption lines are accessible .
all of the predicted galactic fractions @xmath19 at @xmath175 are in between the reported values of @xmath16 for metal lines , @xmath176 ( cowie et al .
1995 ; tytler et al . 1995 ; womble et al .
1996 ) and @xmath177 ( songaila and cowie 1996 ) .
the dependence of @xmath19 on the column density @xmath79 at @xmath178 is shown in fig.3(b ) .
there is an indication of the increase of @xmath19 with @xmath79 ; it would tend to be 1 at @xmath79 around @xmath179 @xmath7 for damped @xmath0 clouds .
there are no observational support yet for the above behaviour of @xmath19 with @xmath79 and will be a test for the two population model .
recently , songaila and cowie ( 1996 ) claim that @xmath180 for @xmath181 @xmath7 and @xmath177 for @xmath182 @xmath7 , whereas the other observations give a much lower value of @xmath16 .
more observations are needed to settle this issue ( see also discussion ) . as we mentioned in 2 , the exact value of @xmath28 , which separates the range of circular velocities into those of minihalos and galactic halos , is yet to be settled from theoretical points of view .
for example ,
even for @xmath1 km s@xmath2 where the virial temperature is @xmath183 k , some fraction of gas in an inner part of minihalos may be able to cool and form stars , because of self - shielding effects against the external uv background radiation .
we thus turn the argument and address the question , as to which range of @xmath28 is allowed in the light of the keck results .
figure 4 shows the galactic fraction of @xmath0 lines as a function of @xmath28 , while @xmath184 km s@xmath2 and @xmath185 km s@xmath2 .
here we note that the change of @xmath28 is applied to the range of integration in eq.(12 ) and the definition of @xmath135 in eq.(13 ) , whereas the range of @xmath9 for the expressions of @xmath79 ( @xmath82 km s@xmath2 for eq.5 and @xmath83 km s@xmath2 for eq.8 ) is unchanged on the theoretical grounds .
for the observed fraction of @xmath186 , the possible value of @xmath28 to reproduce it is 40 to 60 km s@xmath2 .
this is rather insensitive to the specific assumptions on the bias parameter @xmath130 , cosmological parameters ( @xmath187 ) , and the limiting circular velocities ( @xmath188 ) .
however , to set a more robust constraint on the range of circular velocities or masses for mini / galactic halos , further observational determinations of the fraction of @xmath0 lines associated with metal lines are needed .
several recent observations have also measured the metal abundances of absorption lines at high redshifts ( cowie et al .
1995 ; womble et al .
the presence of heavy elements implies that interstellar gas confined in dark halos has been processed by star formation and subsequent chemical evolution .
while there are some alternative possibilities for enriching gas one representative idea is that igm as a whole has already been processed on a sub - galactic scale _ prior to _ formation of collapsed galactic halos ( songaila & cowie 1996 ) , we present here the hypothesis that enrichment occurs _
after _ halo formation .
we postulate that gas within halos has been cooled quickly enough to fragment into stars , and being enriched . in our model , this is applied to the gas in galactic - sized halos with @xmath189 , where cooling time is well less than dynamical time ( dekel & silk 1986 ; efstathiou 1992 ) . here , we attempt to estimate the possible metal abundances of gas in these halos .
suppose that a halo with @xmath9 at the redshift @xmath137 is enriched with an average metallicity @xmath190 . here
the dependences on @xmath9 and @xmath137 arise from the fact that the halos with different @xmath9 collapse and form at different epochs , say @xmath191 , leading to different epochs of star formation , say @xmath192 , and that the subsequent chemical evolution starting at @xmath192 yields the metals at the redshift @xmath137 .
thus an ensemble of these halos with the metallicity @xmath190 contribute to the metal abundance of @xmath0 lines . then since the probability for a sightline to encounter galactic halos in the velocity range of @xmath189 is given by eq.(12 ) , we arrive at the average metallicity of @xmath0 lines with h i column density larger than @xmath79 : @xmath193 the actual expression for @xmath190 in each halo is admittedly not straightforward to derive and very model - dependent .
physics involved in it are , e.g. merging history of a population of dark halos , gasdynamical and thermal processes of gas confined in these halos , and history of star formation . while the extensive and sophisticated approaches of incorporating these processes exist ( lacey & cole 1993 ; kauffmann et al .
1993 ) , we adopt a more simplified approach which we believe includes the essential properties of the system to deduce the characteristic values of metal abundance .
first , we assume that the formation redshift @xmath191 of the halo with mass @xmath99 is represented by the collapse epoch of a rms , 1 @xmath194 linear density contrast , i.e. , if @xmath195 denotes the growth rate of density perturbation , @xmath191 is derived from @xmath196 , where @xmath99 is related to @xmath9 from eq.(10 ) . according to lacey & cole ( 1993 ) on the basis of their extensive analysis of merging halos
, this scaling relation appears to hold for the typical formation epoch of the _ main parent _ halo which had half or more of the present mass @xmath99 ( see their fig.9 and fig.10 ) .
realistically the halo formation epoch for a given @xmath99 is distributed with some probability arising from different amplitudes of density perturbations , but it can be characterised by the collapse epoch of a typical , 1 @xmath194 perturbation .
second , gas in a virialized , galactic - sized halo must cool quickly ( within a dynamical time ) to ensure star formation , so that @xmath192 , the epoch of star formation , is assumed to be equal to @xmath191 .
this is reasonable , because the cooling time in the galactic scales is less than @xmath197 yr , which is a negligible time span much less than the hubble time in the cosmological context .
third , concerning the subsequent chemical evolution , we adopt the observed age - metallicity relation of metal - poor halo stars in our galaxy ( schuster & nissen 1989 ) , on the grounds that typical sightlines to produce the @xmath0 lines with low h i column density penetrate the halo gas , from which old halo stars may have formed .
the average enrichment rate @xmath198 derived by schuster & nissen ( 1989 ) is of the order of @xmath199 gyr@xmath2 for their sample of halo stars .
this value is also close to the value from the chemical - evolution model of the halo by pagel ( 1989 ) .
therefore , denoting the age of the universe at the redshift @xmath137 is @xmath200 , the metallicity @xmath201 at @xmath137 is derived as , @xmath202 where following the above arguments , the epoch of star formation , @xmath203 , depends on @xmath9 and @xmath137 .
we assume here that at the outer parts of the halo , which concerns us here for the @xmath0 forest lines , the gradient of metallcitiy is negligible .
this is a reasonable assumption because there are indications that star formation in the outer halo of our galaxy has been slow and , therefore , the chemical evolution ( matteucci and franois 1992 ) .
if the variation in star formation rate is caused by the variation in the gas collapse time , then this variation is expected to be small in the outer regions of the halo .
this assumption is consistent with the observation that the average metallicity is almost constant for @xmath0 lines with @xmath97 @xmath7 ( cowie 1995 , songaila and cowie 1996 ) .
figure 5(a ) shows the evolution of the mean abundance for absorption lines with @xmath204 @xmath7 ( solid and dotted lines ) .
it is found that for @xmath205 , the metal abundance significantly increases by more than two orders of magnitude . at @xmath175 ,
it is approximately in agreement with the keck results of [ c / h]@xmath206 .
we remark that the adopted age - metallicity relation of a constant enrichment rate overestimates the halo abundance at @xmath4 : the relation should be applied to stars with metallicity less than @xmath207 of solar , which is the criterion to separate halos stars from disk stars ( schuster & nissen 1989 ) . the subsequent formation of halo stars
is stopped , and most of the enriched halo gas would fall into the disk the hypothesis supported for solving the so - called g - dwarf problem in the solar neighborhood ( e.g. yoshii et al .
thus we assume @xmath208 as an upper limit for abundance of absorption lines in fig.5(a ) , but this does not change our result in the light of the keck results at @xmath209 .
we have also investigated the case when the epoch of star formation @xmath210 is constant , being independent of the mass scale of halos as was done by timmes et al .
( 1995 ) for damped @xmath0 systems .
we find that the evolution of @xmath211 based on our prescription of star formation beginning at the collapse epoch of 1 @xmath194 density contrasts is essentially the same as the case when @xmath212 gyr . also shown in fig.5(a ) with filled circles
are the recent findings of lu , sargent & barlow ( 1996 ) for the abundances ( [ fe / h ] ) of damped @xmath0 systems observed with the keck .
intriguingly , the median values of abundances in such systems are well reproduced by the current models .
this may support the hypothesis , first proposed by lanzetta , wolfe & turnshek ( 1995a ) , that damped @xmath0 systems could represent a spheroidal component of galaxies ( see discussion in 4.2 ) .
the dependence of @xmath211 on the value of @xmath28 is shown in fig.5(b ) .
in contrast with the galactic fraction @xmath19 of @xmath0 lines , @xmath211 turns out to be sensitive to both the bias parameter @xmath130 and the cosmological parameters ( @xmath187 ) at large values of @xmath28 .
it is systematically lower for a larger bias and/or lower - density universe .
this is caused by delayed star - formation epoch @xmath210 , leading to less time for enriching gas .
we have shown that the observed fraction of @xmath0 absorption lines associated with metal lines can be understood in terms of photoionized gas confined in galactic - sized halos with @xmath213 km s@xmath2 , whose number density follows the press - schechter model of hierarchical structure formation . the derived average abundance of heavy elements in absorption systems , combined with the characteristic epochs of star formation and the observed age - metallicity relation in our galactic halo , appears to be reasonably in agreement with the observed abundance at @xmath214 .
the current model also suggests that both the fraction of @xmath0 lines associated with galactic halos and the metallicity in absorption systems are expected to increase with decreasing redshift ( fig.3a and fig.5a ) .
while the explicit dependences of these quantities on redshift are yet to be settled ( see songaila & cowie 1996 ) , we note the findings of steidel ( 1990 ) that the number density of c iv lines increases with decreasing redshift . in our model , this can be explained from the above predicted properties of absorption systems in the following way . in photoionization equilibrium ,
the ratio between the column densities of c iv and h i is a function of the ionization parameter @xmath215 , which is the ratio of the densities of ionizing photons and particles , and of the metallicity @xmath201 ; using the photoionization code of cloudy ( ferland 1993 ) , we find that for @xmath216 to @xmath217 , and for @xmath218 to 0.5 , the ratio @xmath219 is well described as , @xmath220 then , suppose that the number density of @xmath0 lines per unit h i column density and redshift , @xmath221 , is represented by a power law form , @xmath222 where @xmath223 ( sargent et al . 1989 ) and @xmath224 ( lu et al .
1991 ) , and that a fraction @xmath19 of these @xmath0 lines are associated with c iv lines , which may be written as , @xmath225 from fig .
3(a ) , we obtain @xmath226 at @xmath209 .
suppose also that the mean metallicity of absorption lines increases in time following @xmath227 .
then eq.(17 ) and eq.(18 ) give the column density distribution of c iv lines as , @xmath228 where @xmath229 describes the redshift evolution of c iv lines .
it is known from observations that the @xmath230 index is the same for both h i and c iv lines ( cowie et al .
1995 ; bergeron et al .
1994 ) , indicating that the simple equation ( 17 ) holds well .
steidel ( 1990 ) has shown that the index @xmath231 is negative , of order @xmath232 .
this means that the number of c iv lines increases in time . in our model
, this can be due partly to the increase of the galactic fraction @xmath19 in time and to the increase of the metallicity @xmath211 in time .
then , for the set of power - law indices , @xmath233 , @xmath234 , @xmath235 , and @xmath236 , we arrive at @xmath237 , thereby indicating , @xmath238 it is worth noting that the enrichment rate of absorption lines shown in fig.5(a ) is well approximated by this equation .
the redshift evolution for c iv lines used here is , strictly speaking , for high c iv column density lines
. it will be interesting to see whether observations in the near future show a similar redshift evolution for small c iv column density lines . therefore , the derived redshift evolutions of both the fraction of @xmath0 lines associated with galactic halos ( fig.3a ) and the metallicity in absorption systems ( fig.5a ) are approximately consistent with the observed scaling laws for @xmath0 and c iv lines .
the abundances of metals have also been measured in damped @xmath0 systems , which are believed to be protogalaxies .
smith et al .
( 1996 ) and pettini et al .
( 1995 ) have recently reported , by measuring the zn lines as a probe of metallicity , that the typical zn abundance at @xmath239 is @xmath240 of solar , but there is a considerable scatter in the zn abundance , spanning more than two orders of magnitude . using high quality spectra obtained with the keck telescope , lu , sargent & barlow ( 1996 ) have also derived a large scatter of abundances , about a factor of 30 , with @xmath79-weighted mean abundance of 0.028 solar between @xmath241 . in the context of the present model
, such a large scatter may be caused by the different formation epoch of halos , @xmath191 , for a given mass arising from the different amplitudes of primordial density fluctuations and different merging history ( lacey & cole 1993 ) .
also , for the adopted age - metallicity relationship of halo stars , there is a scatter in the ages of the stars of @xmath242 gyr at a given abundance , leading to a scatter of abundances more than an order of magnitude at a given age ( schuster & nissen 1989 ) .
it is worth noting that the evolution of the mean abundance is reproduced from the formation epoch of a typical 1 @xmath194 fluctuation and the mean age - metallicity relationship of halo stars in our galaxy ( fig.5a ) .
this suggests that damped @xmath0 systems represent a metal - poor halo component .
also , songaila and cowie ( 1996 ) found that the ratio si / c is about three times solar , similar to the abundance ratio of galactic halo stars .
the two population model , as presented here , clearly predicts ( a ) the fraction of @xmath0 lines with associated metal lines as a function of redshift and ( b ) the redshift evolution of the average metallicity .
the observations , however , have not yet converged to any firm conclusion and it is too early to rule out other possibilities . as noted in songaila and cowie ( 1996 ) , it will be very difficult to determine the fraction of metal enriched @xmath0 clouds below @xmath243 @xmath7 , because this will mean a sensitivity limit of @xmath244 @xmath7 , which is difficult at present .
they claim to have found associated c iv lines in @xmath245 of clouds with @xmath181 @xmath7 and @xmath246 of clouds with @xmath182 @xmath7 . if confirmed , such a large fraction of enriched clouds will be difficult to explain with the two population model . as emphasized by songaila and cowie ( 1996 ) , the alternative hypothesis of an enriched igm
will then be more appropriate .
we briefly note here that a problem also prevails at low redshifts about the interpretation of the fraction of @xmath0 lines with associated metal lines .
although the association of some of the low column density @xmath0 lines with galaxies seems sure , it is still uncertain as to how physical the association is ( see le brun 1996 for a discussion ) .
it is possible that instead of being directly associated with galactic halos ( as in , for example , mo and morris 1994 ) , the @xmath0 lines trace some structure related to formation of galaxies ( morris and van den bergh 1994 , petitjean 1995 ) .
we note here the possibility , as shown in several recent numerical simulations ( miralda - escud 1995 , hernquist 1995 ) that very low column density @xmath0 lines may arise in mini pancakes and not minihalos . in this case
, the density profile will be much different and our results may not hold .
however , we note that mini pancakes can account for @xmath0 lines with @xmath243 @xmath7 ( miralda - escud 1995 ) and lines with higher h i column density come from the intersecting points of pancakes , and correspond to minihalos or galactic halos of our model . therefore , our results are most certain to hold for @xmath247 @xmath7 . in any case , as explained above , the fraction of @xmath0 lines with associated metal lines for @xmath243 @xmath7 will be difficult to determine in the near future , and our model will not be testable for such values of h i column densities .
incidentally , the case for an enriched igm was described by silk ( 1987 ) much before the discovery of associated metal lines in @xmath0 systems .
they discussed the possibility of enriching the igm by galactic winds from dwarf galaxies .
the detail of such a scenario , however , remains to be worked out .
pregalactic objects such as population iii stars might also enrich the igm at high redshifts , but again no detail models exist . in the face of such interesting but still poorly understood scenarios
, it therefore seems reasonable to pursue a modified version of the minihalo model , such as the two population model presented here .
we have shown that a fraction of @xmath13 of @xmath0 forest lines with associated metal lines can be understood in the framework of hierarchical structure formation model , assuming that halos with velocity dispersion @xmath248 km / s do not produce stars and heavy elements .
the model predicts that the fraction of clouds with associated metal lines increases slowly with the h i column density .
we have shown that the results are robust and do not depend strongly on the assumptions of the limiting values of velocity dispersion .
we have further used the age - metallicity relation of galactic halo stars to predict the evolution of the average metallicity of the associated metal line systems . in particular , we find that the average metallicity at @xmath239 should be about a hundredth of solar .
we thank t. padmanabhan , r. srianand , and e. j. wampler for stimulating discussions and helpful comments .
we are also grateful to the referee for his constructive suggestions for the improvement of this paper .
this work has been supported in part by the grand - in - aid for scientific research ( 08640318 ) of the ministry of education , science , and culture in japan .
we describe below how to obtain the h i column density as a function of the impact parameter @xmath64 for @xmath0-absorbing gas clouds in galactic halos .
the probability @xmath249 that a sightline in the range of @xmath250 encounters the clouds confined in the ambient hot gas is given as @xmath251 , where @xmath252 is the number density of clouds and @xmath253 is the cross section of each cloud .
under the assumption that each sightline encounters one cloud in average , @xmath254 is normalized as @xmath255 .
then , if each cloud has its own column density @xmath256 as a function of the galactocentric distance @xmath49 , the average column density @xmath257 at the impact parameter @xmath64 may be written as , @xmath258 to obtain the possible expression for @xmath252 , we assume that the whole area of a galactic halo projected onto the sky is in average covered by ensemble of clouds without overlapping .
this is equivalent to @xmath255 as described above .
then , the average number density of clouds within @xmath64 , denoted as @xmath259 , may be approximately given as @xmath260 , and this yields the number density @xmath252 at each @xmath64 as @xmath261 .
equation ( a1 ) is then reduced to @xmath262 here we note that @xmath263 at @xmath52 where h i column densities of @xmath0 absorption lines are lower than @xmath264 @xmath7 in galactic halos . substituting this into eq.(a2 ) , we obtain @xmath265 where we introduce the variable @xmath266 by @xmath267 in the second equation .
the integration in eq.(a3 ) yields 0.91 , and this can be regarded as @xmath268 in view of the other uncertainties .
this allows us to adopt the approximate equality as @xmath269 .
bahcall , j.n .
, et al . 1993 , , 87 , 1 barcons , x. , lanzetta , k.m . , & webb , j.k . 1995 , , 376 , 321 black , j. 1981 , , 197 , 553 bowen , d.v . , blades , j.c . , & pettini , m. 1996 , , 464 , 141 chiba , m. & nath , b.b .
1994 , , 436 , 618 cowie , l.l . ,
songaila , a. , kim , t.s . ,
& hu , e.m . 1995 , , 109 , 1522 cristiani , s. , dodorico , s. , fontana , a. , giallongo , e. , & savaglio , s. 1995 , , 273 , 1016 dekel , a. & silk , j. 1986 , , 303 , 39 efstathiou , g. 1992 , , 256 , 43p fall , s.m . &
rees , m.j .
1985 , , 298 , 18 ferland , g.j .
1993 , university of kentucky department of physics and astronomy internal report fernndez - soto , a. , et al .
1996 , , 460 , l85 hernquist , l. , katz , n. , weinberg , d.h .
& miralda - escud , j. 1995 , submitted to ( astro - ph/9509105 ) ikeuchi , s. 1986 , , 118 , 509 kauffmann , g. , white , s.d.m .
, & guiderdoni , b. 1993 , , 264 , 201 kulkarni , v. , & fall , s.m .
1993 , , 413 , l63 lacey , c. & cole , s. 1993 , , 262 , 627 lanzetta , k.m . ,
wolfe , a.m. , & turnshek , d.a .
1995a , , 440 , 435 lanzetta , k.m . , bowen , d.b .
, tytler , d. , & webb , j.k .
1995b , , 442 , 538 le brun , v. , bergeron , j. , boisse , p. 1996
, , 306 , 691 lu , l. , wolfe , a.m. , & turnshek , d.a .
1991 , , 367 , 19 lu , l. , sargent , w.l.w . , &
barlow , t.a .
1996 , preprint ( astro - ph/9606044 ) madau , p. 1992
, , 389 , l1 maloney , p. 1992
, , 398 , l89 matteucci , f. , & franois , p. 1992
, , 262 , l1 miralda - escud , j. , cen , r. , ostriker , j.p . , & rauch , m. 1995 , submitted to ( astro - ph/9511013 ) mo , h.j .
, miralda - escud , j. , & rees , m.j .
1993 , , 264 , 705 mo , h.j . , & morris , s.l .
1994 , , 269 , 52 mo , h.j .
1994 , , 269 , l49 morris , s.l . ,
weymann , r.j . , dressler , a. , mccarthy , p.j . ,
smith , b.a . , terrile , r. , giovanelli , r. , & irwin , m. 1993 , , 419 , 524 morris , s.l .
, & van den bergh , s. 1994 , , 427 , 696 moore , b. 1994 , , 370 , 629 ostriker , j.p . & steinhardt , p.j .
1995 , , 377 , 600 padmanabhan , t. 1993 , structure formation in the universe ( cambridge : cambridge univ . press ) pagel , b.e.j .
1989 , rev .
mexicana astron .
astrofis . ,
18 , 161 peebles , p.j.e .
1980 , the large - scale structure of the universe ( princeton : princeton univ .
press ) peterson , b.a .
1978 , in the large - scale structure of the universe , eds .
longair , j. einasto ( dordrecht : reidel ) , p. 389 petitjean , p. , webb , j.k . ,
rauch , m. , carswell , r.f . , & lanzetta , k. 1993 , , 262 , 499 petitjean , p. , mcket , j. , & kates , r.e .
1995 , , 295 , l9 pettini , m. , king , d.l . ,
smith , l.j . , & hunstead , r.w .
1995 , in qso absorption lines , ed .
g. meylan ( berlin : springer - verlag ) , p. 71
postman , m. & geller , m.j .
1984 , , 281 , 95 press , w.h . , schechter ,
, , 187 , 425 rauch , m. , carswell , r.f . ,
chaffee , f.h . ,
foltz , c.b . ,
webb , j.k .
, weymann , r.j . , bechtold , j. , & green , r.f .
1992 , , 390 , 387 rauch , m. , sargent , w.l.w . ,
womble , d.s . , &
barlow , t.a .
1996 , , 467 , l5 rees , m.j .
1986 , , 218 , 25p salpeter , e. & hoffmann , g.l .
1995 , , 441 , 51 sargent , w.l.w . ,
young , p.j . ,
boksenberg , a. , & tytler , d. 1980 , , 42 , 41 sargent , w.l.w . ,
steidel , c.c . , & boksenberg , a. 1989 , , 69 , 703 schuster , w.j . & nissen , p.e .
1989 , , 222 , 69 silk , j. , wyse , r.f.g . , & shields , g. a. 1987 , , 322 , l59 smith , l.j . ,
pettini , m. , king , d.l . , & hunstead , r.w .
1996 , in the impact of stellar physics on galaxy evolution , eds . c. leitherer ,
u. f .- v .
alvensleben , j. huchra , in press songaila , a. , & cowie , l.l .
1996 , , in press ( astro - ph/9605102 ) steidel , c.c .
1990 , , 72 , 1 suto , y. 1993 , progress of theoretical physics , 90 , 1173 thoul , a.a .
& weinberg , d.h .
1996 , , 465 , 608 timmes , f.x . ,
lauroesch , j.t . , & truran , j.w .
1995 , , 451 , 468 tytler , d. , fan , x .-
m . , burles , s. , cottrell , l. , davis , c. , kirkman , d. , & zuo , l. 1995 , , in qso absorption lines , ed .
g. meylan ( berlin : springer - verlag ) , p. 289
womble , d.s , sargent , w.l.w . , & lyons , r.s .
1996 , in cold gas at high redshift , eds .
m. bremer , h. rottgering , p. van der werf , c. carilli ( kluwer ) , in press yoshii , y. & arimoto , n. 1987 , , 188 , 13 yoshii , y. , tsujimoto , t. , & nomoto , k. 1996 , , 462 , 266 | we investigate the hypothesis that @xmath0 absorption lines arise in two populations of halos minihalos of small circular velocity ( @xmath1 km s@xmath2 ) in which star formation and metal production are inhibited by photoionization of the uv background radiation , and large galactic halos ( @xmath3 km s@xmath2 ) which possess stars and metals .
based on the model of @xmath0-absorbing gas confined in both populations of halos , we attempt to explain the recent observations of ( 1 ) associations of visible galaxies with @xmath0 lines at low redshifts @xmath4 , and ( 2 ) metal lines associated with a non - negligible fraction of low h i column density @xmath0 lines at @xmath5 . for galactic halos , we find that photoionized gas clouds confined in the pressure of ambient hot gas can produce @xmath0 absorptions with h i column density as low as @xmath6 @xmath7 , and that the impact parameter of a sightline for such absorptions matches well with the observed radius of gaseous envelope in a typical luminous galaxy . using the press - schechter prescription for the mass function of halos
, we also show that the fraction of @xmath0 lines with associated metal lines can be understood in terms of the fraction of @xmath0 absorbers that are associated with galactic halos .
in particular , the reported fraction of @xmath8 at @xmath5 is reproduced when the boundary value of @xmath9 to separate mini or galactic halos is @xmath10 km s@xmath2 , which is consistent with the theoretical prediction of galaxy formation under photoionization .
the average metallicity of both @xmath0 forest and damped @xmath0 systems at @xmath5 is explained in terms of the model of halo - formation history combined with the age - metallicity relationship of galactic halo stars .
possible methods to test this hypothesis and the other alternative scenarios are also discussed .
21j_-21 21_-21 |
this is a continuation of the paper @xcite .
we showed there that the trigonometric wzw model is factorised into the orbifold wzw models . using this result
, we show in the present article that the trigonometric wzw model is indeed the degenerate twisted wzw models on elliptic curves defined in @xcite .
more precisely , we prove that there are locally free sheaves over the partially compactified family of elliptic curves , the fibre of which are the space of conformal blocks or the space of conformal coinvariants of the twisted wzw model at a generic point and those of the trigonometric wzw model at the discriminant locus , when all inserted modules are either weyl modules ( proposition [ prop : coherence ] ) or integrable highest weight modules ( theorem [ thm : loc - free ] ) . since the elliptic @xmath0 matrix ( cf .
@xcite , @xcite ) describing the elliptic kz equations degenerates to the trigonometric @xmath0 matrix , the above fact is naturally expected , though rigorous proof requires careful algebro - geometric arguments as in @xcite .
the paper is organised as follows .
after reviewing the twisted wzw model on elliptic curves in [ sec : twwzw - ell ] to recall basic notions and notations , we define family of elliptic curves with a singular fibre and a twisted lie algebra bundle over it in [ sec : family ] . in [ sec : cc , cb ] main objects of this paper , the sheaves of conformal coinvariants and conformal blocks , are defined and their coherence is proved . in particular when all the modules inserted to the curve are weyl modules , they are locally free . to prove the locally freeness of the sheaf of conformal coinvariants for integrable highest weight modules , we examine its behaviour at the discriminant locus . in this case the factorisation theorem , theorem 7 of @xcite , is refined in
[ sec : sheaf - tri / orb ] .
the proof of locally freeness in [ sec : loc - free ] follows the strategy of @xcite and @xcite .
we use the following notations besides other ordinary conventions in mathematics .
* @xmath1 , @xmath2 : fixed integers .
@xmath3 will be the matrix size and @xmath4 will be the number of the marked points on a curve .
* @xmath5 : the cyclic group of order @xmath1 .
* when @xmath6 is an algebraic variety , @xmath7 denotes the structure sheaf of @xmath6 .
when @xmath8 is a point on @xmath6 and @xmath9 is an @xmath7-sheaf , @xmath10 denotes the stalk of @xmath9 at @xmath8 .
@xmath11 is the maximal ideal of the local ring @xmath12 .
@xmath13 , @xmath14 are the fibre of @xmath9 at @xmath8 and the @xmath11-adic completion of @xmath10 , respectively .
* we shall use the same symbol for a vector bundle and for a locally free @xmath7-module consisting of its local holomorphic sections .
in this section we briefly review the twisted wzw model on elliptic curves .
see @xcite for details .
we fix an invariant inner product of @xmath15 by @xmath16 define matrices @xmath17 and @xmath18 by @xmath19 where @xmath20
. then we have @xmath21 and @xmath22 .
let @xmath23 be the elliptic curve with modulus @xmath24 : @xmath25 .
we define a lie algebra bundle @xmath26 with fibre @xmath27 over @xmath28 by @xmath29 where the equivalence relations @xmath30 are defined by @xmath31 let @xmath32 , which satisfies @xmath33 global meromorphic sections of @xmath26 are linear combinations of @xmath34 ( @xmath35 , @xmath36 ) , where @xmath37 is a meromorphic function with quasi - periodicity , @xmath38 for each point @xmath8 on @xmath28 , we define a lie algebra , @xmath39 where @xmath40 is the sheaf of meromorphic functions on @xmath28 and @xmath41 means the completion of the stalk at @xmath8 with respect to the maximal ideal @xmath11 of @xmath42 , the stalk of the structure sheaf .
the lie algebra @xmath43 is ( non - canonically ) isomorphic to the loop lie algebra @xmath44 , where @xmath45 is the coordinate of @xmath8 .
the subspace @xmath46 \label{def : loop+}\ ] ] of @xmath43 is a lie subalgebra .
let us fix mutually distinct points @xmath47 on @xmath28 whose coordinates are @xmath48 and put @xmath49 .
we shall also regard @xmath50 as a divisor on @xmath28 ( i.e. , @xmath51 ) .
the lie algebra @xmath52 has a 2-cocycle defined by @xmath53 where @xmath54 , @xmath55 is the residue at @xmath56 and @xmath57 is the exterior derivation .
( the symbol `` @xmath58 '' stands for `` cocycle defining the affine lie algebra '' . )
we denote the central extension of @xmath59 with respect to this cocycle by @xmath60 : @xmath61 where @xmath62 is a central element .
explicitly the bracket of @xmath60 is represented as @xmath63 = ( [ a_i , b_i]^\circ)_{i=1}^l \oplus { { { \mathrm c}_{\mathrm a}}}(a , b ) { { \hat k}}\quad \text{for $ a , b\in{{\mathfrak g}}^d$ , } \label{def : aff - alg - str}\ ] ] where @xmath64^\circ$ ] are the natural bracket in @xmath65 .
the lie algebra @xmath66 for a point @xmath8 is nothing but the affine lie algebra @xmath67 of type @xmath68 ( a central extension of the loop algebra @xmath69 ) .
the affine lie algebra @xmath70 can be regarded as a subalgebra of @xmath60 .
the subalgebra @xmath71 of @xmath65 ( cf . ) can be also regarded as a subalgebra of @xmath70 and @xmath60 .
let @xmath72 be the space of global meromorphic sections of @xmath26 which are holomorphic on @xmath28 except at @xmath50 : @xmath73 the residue theorem implies that we can regard @xmath72 as a lie subalgebra of @xmath60 by mapping an element of @xmath72 to its germs at @xmath56 s .
[ def : cc , cb : ell ] the space of _ conformal coinvariants _ @xmath74 and that of _ conformal blocks _ @xmath75 over @xmath28 associated to @xmath70-modules @xmath76 with the same level @xmath77 are defined by @xmath78 where @xmath79 .
the module @xmath76 is referred to as `` a module inserted at the point @xmath56 '' .
in this section we construct a family of elliptic curves @xmath80 with a singular fibre and its covering @xmath81 .
a twisted lie algebra bundle @xmath26 over the generic fibres of @xmath80 is one of the main representation theoretical data in the twisted wzw model but it does not directly extends to the singular fibre . hence we pull it back to @xmath81 and trivialise it .
the sections of @xmath26 on the singular fibre is understood as sections of the trivial bundle on @xmath81 invariant under the action of the covering transformation group .
the construction of @xmath81 is almost the same as that of the analytic fibre space of elliptic curves in @xcite .
we use @xmath1 patches @xmath82 ( @xmath83 ) : @xmath84 we denote @xmath85 by @xmath86 .
the universal curve @xmath81 is defined by @xmath87 where the equivalence relation @xmath88 is defined by @xmath89 we have an analytic fibre space @xmath90 over @xmath91 .
the fibre over @xmath92 is an elliptic curve @xmath93 and the fibre over @xmath94 is singular with ordinary double points @xmath95 ( @xmath83 ) .
let @xmath96 be the cyclic group of order @xmath1 .
the group @xmath97 acts on @xmath81 from the right as follows : it is enough to define the action of the generators @xmath98 and @xmath99 of @xmath100 . @xmath101
the universal curve @xmath80 is defined set - theoretically as the quotient space of @xmath81 by this action : @xmath102 the canonical projection to @xmath103 is denoted by @xmath104 .
the fibre @xmath105 is the singular curve with one ordinary double point . -covering . ] the family of @xmath2-pointed elliptic curves and its covering are defined by @xmath106 here the base space @xmath107 is the fibre product of @xmath80 s without diagonals : @xmath108 \text{\ for any k. } \ } , \end{split } \label{def : s}\ ] ] where @xmath56 is the point of the @xmath109-th @xmath80 in the fibre product with @xmath110 .
we exclude the degeneration of the types @xmath111 and @xmath112 ( node ) .
we denote the canonical projections @xmath113 , @xmath114 and @xmath115 by @xmath116 , @xmath117 and @xmath118 respectively .
the fibres of @xmath119 and @xmath120 over @xmath121 are singular curves .
the section @xmath122 of @xmath113 is defined by @xmath123 we denote the divisor @xmath124 + \cdots + [ q_l(s)]$ ] by @xmath50 .
the group @xmath100 acts on @xmath119 naturally as covering transformation and on @xmath125 by @xmath126 the twisted lie algebra bundle @xmath127 on @xmath128 is defined as the associated bundle to the @xmath100-principal bundle @xmath129 : @xmath130 it is obvious that the restriction of @xmath127 to a fibre of @xmath120 at a point @xmath131 ( @xmath132 ) is the bundle @xmath26 on the elliptic curve @xmath133 defined by
. sheaf version of affine lie algebras @xmath43 , @xmath134 , @xmath59 and @xmath60 ( cf . , , ) are @xmath135-lie algebras defined by @xmath136 since we assume that sections @xmath56 do not touch the singular point of the singular fibre , the definitions are the same as those for the non - singular case , ( 3.11 ) of @xcite .
the central extensions of @xmath137 and @xmath138 are defined by the cocycle with the coefficients in @xmath135 : @xmath139 the lie subalgebra of meromorphic sections @xmath140 , , would be replaced by @xmath141 if there were no singularity , as was the case in @xcite .
taking the singular fibre into account , we modify this naive definition as follows : @xmath142 where @xmath143 is the @xmath100-orbit of the divisor @xmath50 and @xmath144 denotes the @xmath100-invariant section of the equivariant locally free sheaf @xmath145 . in other words ,
a @xmath125-valued meromorphic function @xmath146 ( @xmath147 , @xmath148 ) belongs to @xmath149 if and only if it satisfies @xmath150 where @xmath151 ( @xmath152 ) is the left action of the generators of @xmath100 on the fibre of @xmath81 defined by @xmath153 .
( see . )
our construction is so explicit that we have an explicit basis of @xmath149 . in @xcite meromorphic functions @xmath154
( @xmath35 , @xmath155 , @xmath156 ) characterised by the following properties were introduced : * additive quasi - periodicity : @xmath157 , @xmath158 ; * as a function of @xmath156 , @xmath154 has a simple pole with residue @xmath159 at @xmath160 .
let us denote this function by @xmath161 .
( the superscript `` add '' stands for `` additive '' . )
let us rewrite it to a multiplicatively quasi - periodic function @xmath162 as follows : @xmath163 where @xmath164 is the standard infinite product symbol .
the function @xmath162 is related to @xmath165 by @xmath166 when @xmath132 , that is , we replaced the arguments of @xmath165 by @xmath167 , @xmath168 and used the product formula for the theta function .
when @xmath94 , @xmath169 becomes a rational function of @xmath170 : @xmath171 the important property of @xmath169 is that it inherits the quasi - periodicity of @xmath165 : @xmath172 we define the function @xmath173 ( @xmath174 ) on @xmath119 in terms of @xmath169 : for @xmath175 ( cf . ) , @xmath132 , @xmath176 here @xmath177 and we fix an index @xmath178 to express @xmath56 as a point @xmath179 in @xmath180 .
( the function @xmath181 is determined up to this choice . )
this function is extended to the points with @xmath94 .
( for example , @xmath182 is the rational function of @xmath183 when @xmath56 is represented as a point @xmath184 in @xmath185 . )
the main properties of this function are * quasi - periodicity with respect to @xmath100-action : @xmath186 * all poles are simple and located at @xmath56 modulo @xmath100 action .
it is easy to see that any section of @xmath149 is a linear combination of @xmath187 s and their derivatives along the fibre .
[ lem : gd = gout+regular ] @xmath188 .
the singular part of an element of @xmath189 can be expressed by a linear combination of @xmath187 and derivatives in a unique way .
subtracting such combination which belongs to @xmath149 , we end up with a regular element of @xmath190 .
in this section we introduce the sheaf @xmath191 of conformal coinvariants and the sheaf @xmath192 of conformal blocks and show their basic properties .
definitions of @xmath191 and @xmath192 are literally the same as those for the non - singular case , definition 3.3 of @xcite .
[ def : scc - scb ] for any @xmath189-module @xmath193 of level @xmath194 ( i.e. , @xmath62 acts as @xmath195 ) , we define the _ sheaf @xmath196 of conformal coinvariants _ and the _ sheaf @xmath197 of conformal blocks _ by @xmath198 we can regard @xmath199 as a covariant right exact functor from the category of @xmath189-modules to that of @xmath135-modules and similarly @xmath200 as a contravariant left exact functor . the goal of this paper is to prove that @xmath191 and @xmath192 are locally free . since @xmath192 is the dual of @xmath191 , we mainly discuss about @xmath191 and briefly mention on @xmath192 when it is necessary .
we assume that the @xmath189-module @xmath193 are of the following type : @xmath201 where each @xmath76 is a quotient of a @xmath67-weyl module @xmath202 of level @xmath194 for a finite dimensional irreducible @xmath125-module @xmath203 .
( see 6 of @xcite or
2.4 of @xcite . ) to endow @xmath193 in with the @xmath189-module structure , we need to fix the coordinate of @xmath80 and the trivialisation of @xmath127 , which is irrelevant to the statements of theorems below . in the concrete computations
, we use the coordinates and the trivialisation obtained naturally from the construction in [ sec : family ] .
[ prop : coherence ] @xmath196 is a coherent @xmath135-sheaf . when all @xmath76 s are weyl modules , @xmath196 is ( and hence @xmath197 is ) locally free .
lemma [ lem : gd = gout+regular ] makes it possible to apply the same argument as the proof for the non - singular case , corollary 3.5 of @xcite .
in fact , if each @xmath76 in is a weyl module @xmath204 , @xmath193 is expressed as @xmath205 by the poincar - birkhoff - witt theorem ( @xmath206 ) . here
@xmath207 denotes the universal @xmath135-enveloping algebra . modding out by @xmath208
, we have @xmath209 which means that @xmath196 is locally free , in particular , coherent . if @xmath76 s are quotients of weyl modules , @xmath193 is a quotient of @xmath210 .
hence by the right exactness of the functor @xmath191 , @xmath196 is a quotient of a coherent sheaf , and therefore coherent . in [ sec : loc - free ] we prove the locally freeness of @xmath196 for integrable @xmath76 s , examining the behaviour of @xmath196 at the boundary of the moduli space ( @xmath211 ) carefully .
everything in previous two sections can be restricted on @xmath212 , namely on the configuration space of points on a singular rational curve with one ordinary double point .
( as we have mentioned , the restriction of the functions @xmath181 needs special care . ) hence we can define the corresponding sheaves @xmath196 and @xmath197 which we denote by @xmath213 and @xmath214 .
the subscript `` trig '' is put here because , as we shall see below , there are connections on them expressed in terms of the trigonometric @xmath0 matrix . in the proof of theorem [ thm : loc - free ]
we shall use the sheaf of conformal coinvariants of the orbifold wzw model , @xmath215 on @xmath212 , where @xmath216 for a @xmath217-module @xmath218 ( @xmath219 ) .
we shall recall the definition of the twisted affine algebras @xmath220 and @xmath221 and the details of @xmath222 soon later . here
we only say that @xmath223 is defined exactly in the same way as @xmath213 if we replace the degenerate elliptic curve ( the fibre of @xmath80 at @xmath94 ) by the orbifold @xmath224 .
note that @xmath212 can be regarded as the configuration space of points on @xmath224 .
( see 3 of @xcite . )
let @xmath193 be as in and @xmath225 be a @xmath226-verma module .
then proposition [ prop : coherence ] holds for @xmath227 , @xmath228 , @xmath223 and @xmath229 as well .
in fact , we can prove locally freeness under this assumption .
[ prop : loc - free : trig / orb ] ( i ) the sheaf @xmath213 and the sheaf @xmath214 are locally free @xmath230-sheaves .
\(ii ) the sheaf @xmath215 and the sheaf @xmath231 are locally free @xmath230-sheaves .
\(i ) the proof of the locally freeness for the non - singular case , corollary 5.3 of @xcite is true also in this case : @xmath213 is coherent as shown at the end of [ sec : cc , cb ] and there is a connection and @xmath50-module structure on it , which implies that it is locally free @xmath230-sheaf .
the only difference is that we do not change the curve itself ( the modulus @xmath232 is fixed to @xmath233 ) in the present case , and hence there is nothing corresponding to the connection in the direction of @xmath234 in @xcite .
the connection in the direction of @xmath235 ( @xmath236 is the coordinate of @xmath56 ) is @xmath237)$ ] ( cf .
( 5.14 ) of @xcite ) as is well - known , where @xmath238 is the representation of the virasoro algebra on @xmath76 constructed via the sugawara construction and @xmath239 $ ] is one of the virasoro generator , usually denoted by @xmath240 .
\(ii ) we might proceed as the proof of ( i ) from the beginning but the short cut is to use the result of ( i ) .
the coherence of @xmath241 having been proved , we have only to check that the above connection operators @xmath242 ( @xmath174 ) on @xmath213 also define the flat connection on @xmath223 . what we need to check is * @xmath243 \subset { { { \mathfrak g}}_{\mathrm{out}}}^{\mathrm{orb}}$ ] , * @xmath244 = 0 $ ] on @xmath223 , which is proved in the same way as in the case of the ordinary wzw model , e.g. , lemma 4 of @xcite .
the connection on @xmath213 mentioned in the proof of ( i ) is obtained by the degeneration @xmath245 of the elliptic knizhnik - zamolodchikov connection in @xcite and @xcite .
they are expressed as the first order differential operators on @xmath246 in terms of the _ trigonometric @xmath0 matrix_. in fact , by tracing the argument which leads to the explicit form ( theorem 5.9 in @xcite ) of the connection , we have only to replace the functions @xmath247 there with @xmath248 which is expressed by the rational function of the form on @xmath212 . hence the kz equation for the wzw model on the degenerate elliptic curve is the _ trigonometric kz equation_. [ lem : cctri / orb - fibre ] ( i ) the fibre of @xmath213 at @xmath249 , @xmath250 , is isomorphic to @xmath251 , the space of conformal coinvariants of the trigonometric wzw model for the geometric data corresponding to @xmath252 .
\(ii ) the fibre of @xmath253 at @xmath252 , @xmath254 , is isomorphic to @xmath255 , the space of conformal coinvariants of the orbifold wzw model for the geometric data corresponding to @xmath252 .
see definition 3.2 of @xcite for the definition of @xmath256 and @xmath222 .
we can modify the proof of the corresponding statement for the non - singular case , corollary 5.4 of @xcite .
for example , for the case ( i ) , the isomorphism @xmath257 is a consequence of the existence of the sections @xmath181 and their derivatives ( cf .
the end of [ sec : family ] ) which span both @xmath258 and @xmath259 .
the rest of the proof can be translated to the present case without change .
the proof of ( ii ) is similar . combining proposition [ prop : loc - free : trig / orb ] , lemma [ lem : cctri / orb - fibre ] and theorem 5.1 of @xcite , we have the following isomorphism : @xmath260 where @xmath261 is the @xmath125-module generating @xmath262 ( recall @xmath76 is a quotient of weyl module @xmath204 ) , @xmath263 is the set of its weights , @xmath264 @xmath265 for a verma module @xmath266 of @xmath226 with the highest weight @xmath267 ( cf .
definition 4.1 ( i ) of @xcite ) . in [ sec :
loc - free ] the modules @xmath76 are assumed to be integrable highest weight modules .
chapter 10 of @xcite . ) in this case the above result can be refined .
for this purpose we recall the details of the orbifold wzw model defined in 3 of @xcite .
let us denote the standard coordinate of @xmath268 by @xmath269 .
the cyclic group @xmath270 acts as @xmath271 ( @xmath272 ) and the quotient @xmath273 is an ordinary orbifold .
the definition of the space of conformal coinvariants / blocks of the orbifold wzw model on @xmath274 is almost the same as that on elliptic curves , definition [ def : cc , cb : ell ] , except that we also insert modules to the singular points @xmath233 and @xmath275 .
the lie algebra @xmath72 in is replaced by @xmath276 which consists of @xmath125-valued meromorphic functions @xmath277 on @xmath278 such that : ( 1 ) poles belong to @xmath279 ; ( 2 ) @xmath280 .
accordingly , the module inserted at @xmath233 is the @xmath220-module and the module inserted at @xmath275 is the @xmath221-module , where @xmath281 the cocycles which defines the central extension of @xmath282 and @xmath283 are : @xmath284 for @xmath285 and @xmath286 respectively .
( @xmath287 is the loop algebra part of @xmath226 . ) as etingof showed ( lemma 1.1 of @xcite ) , @xmath220 and @xmath221 are isomorphic to the ordinary affine lie algebra @xmath67 of @xmath68 type .
explicitly the isomorphism @xmath288 is defined by @xmath289 for @xmath290 , where @xmath291 is the matrix unit , @xmath292 , @xmath293 , @xmath294 ( @xmath295 ) are the chevalley generators of @xmath67 and @xmath296 ( @xmath295 ) are coroots of @xmath67 .
6.2 and 7.4 of @xcite . )
since the positive powers of @xmath252 ( cf . )
kill the highest weight vector of the verma module @xmath297 inserted at @xmath275 ( cf.definition 4.1 ( i ) of @xcite ) , we identify @xmath221 and @xmath67 through an isomorphism which is essentially a composition of @xmath298 and the chevalley involution : @xmath299 for @xmath290 . identified through @xmath298 and @xmath300 , the verma modules of @xmath220 and @xmath221 are verma modules of @xmath67 in ordinary sense . in this section , @xmath293 and @xmath294 denote the chevalley generators of @xmath67 identified with the elements in @xmath220 or @xmath221 by means of @xmath298 or @xmath300 . [
prop : ccorb : integrable ] assume that all @xmath76 ( @xmath174 ) are integrable highest weight modules and that @xmath301 ( @xmath302 ) is a verma modules of @xmath217 . then @xmath303 is @xmath233 unless @xmath304 and @xmath305 are dominant integral weights of @xmath67 identified with @xmath220 and @xmath221 . if @xmath304 and @xmath305 are dominant integral weights , @xmath306 where @xmath307 ( @xmath302 ) is the irreducible quotient of @xmath301 . in physics context
, this proposition is a consequence of the propagation of the null field .
see 4 of @xcite .
the author thanks yasuhiko yamada for this comment .
proposition [ prop : ccorb : integrable ] is in sharp contrast to the weyl module case .
see 6 of @xcite .
the following lemma shall be proved later .
[ lem : cc - reduction ] let @xmath308 ( @xmath309 ) be a quotient of the verma module @xmath301 .
suppose @xmath310 is a singular vector of weight @xmath311 which is not a dominant integral weight .
( the weight @xmath311 may possibly be the highest weight @xmath304 . )
then for any @xmath312 and @xmath313 , @xmath314 the same is true for a singular vector @xmath315 .
the first statement of proposition [ prop : ccorb : integrable ] is a consequence of lemma [ lem : cc - reduction ] .
for example , assume @xmath304 is not a dominant integral weight .
let us show @xmath316\cdots x_l[-n_l]|\mu_0\rangle { \otimes}v { \otimes}v_\infty \equiv 0 \mod{{{{\mathfrak g}}_{\mathrm{out}}}^{\mathrm{orb } } } , \label{reduction : not - integrable:1}\ ] ] where @xmath317 \in \hat{{\mathfrak g}}^{(0)}$ ] ( @xmath318 , @xmath319 ) , @xmath320 , @xmath321 and @xmath322 denotes @xmath323 .
( this abbreviation shall be used throughout this paper . )
let @xmath324 be an element of @xmath276 such that @xmath325 for sufficiently large @xmath326 .
( such @xmath327 exists due to the riemann - roch theorem .
it is not difficult to construct such a function concretely .- valued function @xmath277 with poles in @xmath328 , @xmath329 . ] ) then we may replace @xmath330 $ ] by @xmath331 : @xmath332\cdots x_l[-n_l]|\mu_0\rangle { \otimes}v { \otimes}v_\infty \\ = & \rho_0(f_1(t ) ) x_2[-n_2]\cdots x_l[-n_l]|\mu_0\rangle { \otimes}v { \otimes}v_\infty \\ \equiv & - x_2[-n_2 ] \cdots x_l[-n_l]|\mu_0\rangle { \otimes}\\ & { \otimes}\left(\sum_{i=1}^l \rho_i(f_1(t ) ) v { \otimes}v_\infty + v { \otimes}\rho_\infty(f_1(t ) ) v_\infty\right ) \mod{{{{\mathfrak g}}_{\mathrm{out}}}^{\mathrm{orb}}}. \end{split } \label{reduction : temp1}\ ] ] by induction on @xmath333 , the problem is reduced to showing @xmath334 which immediately follows from lemma [ lem : cc - reduction ] . to prove the second statement of proposition [ prop : ccorb : integrable ] , assume that @xmath304 and @xmath305 are dominant integral weight .
then the irreducible quotients of @xmath301 are expressed as @xmath335 see ( 10.4.6 ) of @xcite . therefore to prove , it is enough to show @xmath336 and a similar statement with the indices `` @xmath233 '' and `` @xmath275 '' for @xmath337 , @xmath338 etc .
they are proved as above , namely by the arguments like and , because the weight of the singular vector @xmath339 is not a dominant integral weight .
since @xmath311 is not a dominant integral weights , there is an index @xmath109 ( @xmath340 ) such that @xmath341 is not a non - negative integer . by an easy calculation
, we have @xmath342 for any @xmath343 .
note that the constant @xmath344 never vanishes .
let @xmath345 be an element of @xmath276 such that : ( 1 ) @xmath346 for sufficiently large @xmath326 ; ( 2 ) @xmath347 ( i.e. , @xmath345 has a zero of large order at @xmath348 ) .
such an element can be constructed in the form @xmath349 , where @xmath350 ( @xmath290 ) or @xmath351 ( @xmath352 ) and @xmath353 is a rational function .
hence we can rewrite @xmath354 modulo @xmath276 as follows ( cf .
p.479 of @xcite ) : @xmath355 recall that @xmath356 is locally nilpotent on @xmath357 ( corollary 1.4.6 of @xcite ) .
thus the right hand side of is @xmath233 for large @xmath326 , which completes the proof of the lemma . because of the difference of the sign in and and the fact that @xmath358 is a dynkin automorphism , if @xmath359 is a dominant integral weight for @xmath220 , @xmath360 is a dominant integral weight for @xmath221 and vice versa .
( see . )
[ cor : factor : integrable ] under the same assumption as proposition [ prop : ccorb : integrable ] we have @xmath361 where `` dom
. int . '' means `` dominant integral weight '' .
similarly the decomposition becomes @xmath362 where @xmath363 ( @xmath364 ) .
the main theorem of this paper is proved in this section .
we show that @xmath191 is locally free at the discriminant locus , @xmath365 , provided that all modules inserted are _
integrable highest weight modules_. thus , combining the result in @xcite , we have locally freeness of the sheaf of conformal coinvariants and consequently locally freeness of its dual , the sheaf of conformal blocks . corresponding statement for the weyl modules has been proved in proposition [ prop : coherence ] .
the arguments in this section is parallel to those in 7.3 of @xcite .
we assume the condition for @xmath193 .
moreover we assume that all @xmath76 s are integrable highest weight modules .
in particular , the level @xmath194 is a non - negative integer .
[ thm : loc - free ] the sheaf @xmath196 and hence the sheaf @xmath197 are locally free @xmath366-sheaves .
the rest of the paper is devoted to the proof of this theorem .
the main strategy of the proof is the same as that of @xcite .
see also @xcite , @xcite and @xcite : * since @xmath196 is coherent , it is sufficient to prove that each stalk @xmath367 ( @xmath368 ) is a free @xmath369-module .
* we proved the locally freeness of @xmath196 on the non - singular part @xmath370 in @xcite .
thus we have only to prove the case @xmath371 .
* when @xmath371 , we prove that the stalk of the completion of @xmath196 at @xmath252 is isomorphic to @xmath372 $ ] .
* @xmath367 is a free @xmath369-module because of proposition [ prop : loc - free : trig / orb ] and the faithfully flatness of the completion functor . we define completion of the sheaf @xmath196 along the divisor @xmath212 of @xmath107 by taking completion of each ingredient of the definition .
let @xmath373 be the completion of @xmath135 along @xmath212 : @xmath374 ( @xmath375 is the definining ideal of @xmath212 . ) as an @xmath230-module , it is isomorphic to the ring of formal power series : @xmath376 $ ] .
the completion of @xmath377 is obviously @xmath378 , \label{hatm}\ ] ] where @xmath379 .
the @xmath135-lie algebra @xmath149 acts on @xmath380 naturally as follows : a germ @xmath381 of @xmath149 at @xmath382 is expanded at @xmath383 in terms of the coordinates @xmath384 as @xmath385 where @xmath386 ( @xmath249 ) and @xmath387 note that the periodicity condition implies @xmath388 and hence @xmath389 and @xmath390 are meromorphic ( rational ) function on @xmath278 with quasi - periodicity @xmath391 the poles of which are in the divisor @xmath392 + \ast [ 0 ] + \ast [ \infty ] .
\label{def : tilded0}\ ] ] therefore @xmath393 and @xmath394 , namely the @xmath326-th coefficients of the expansion @xmath395 define a section @xmath396 of @xmath397 with poles at @xmath398 .
the action of @xmath381 on @xmath399 ( @xmath400 , @xmath401 ) is defined by @xmath402 here @xmath238 denotes the usual action of the laurent expansion of @xmath396 at @xmath56 on @xmath76 .
the space of coinvariants of @xmath380 with respect to this action is the completion of @xmath196 : @xmath403 [ lem : completion ] @xmath404 by definition , we have an exact sequence @xmath405 tensoring @xmath373 we obtain an exact sequence @xmath406 \to { { \widehat{{\mathcal m}}}}\to { \mathcal{cc}}({{\mathcal m}}){\otimes}{{\hat{{\mathcal o}}_{s / s_0}}}\to 0 .
\label{completion : goutxm->m->cc}\ ] ] it is sufficient to show that the image of the map @xmath407 \to { { \widehat{{\mathcal m}}}}$ ] is @xmath408 defined by the action .
this is almost trivial since @xmath409 for @xmath410 and @xmath411 is expressed as @xmath412 because of the expansion .
next step is to make a completion of the isomorphism . for this purpose
we need a lemma on verma modules of @xmath220 and @xmath221 . note that the verma module @xmath266 of @xmath226 ( @xmath413 ) is graded by the degree : @xmath414 \cdots x_l[-n_l ] |\mu\rangle , \label{verma : degree}\ ] ] where @xmath415 s are the highest weight vectors of @xmath416 and @xmath317 \in \hat{{\mathfrak g}}^{(*)}$ ] ( @xmath417 , @xmath418 ) .
[ lem : pairing - verma ] ( i ) for any @xmath419 there exists a pairing between the verma modules @xmath420 and @xmath421 : @xmath422 which satisfies @xmath423 and @xmath424u , v\right > + \left < u , { \operatorname{ad}}\beta(x)[-n ] v \right > = 0 , \label{pairing : invariance}\ ] ] for any @xmath425 , @xmath426 , @xmath427 and @xmath428 .
\(ii ) @xmath429 if @xmath430 .
\(iii ) the radical @xmath431 is the largest proper submodule of @xmath432 .
similarly the radical in @xmath421 is the largest proper submodule .
hence the pairing descends to a non - degenerate pairing between the irreducible quotients @xmath433 and @xmath434 .
let @xmath435 be the anti - isomorphism @xmath436 = x { \otimes}t^n \mapsto - x { \otimes}t^n = - x[-n ] \in u\hat{{\mathfrak g}}^{(\infty ) } , \qquad \nu(\hat k ) = \hat k. \label{def : nu : u->u}\ ] ] this induces a linear isomorphism @xmath437 defined by @xmath438 where @xmath439 is the generating vector of the right @xmath221-module @xmath440 , normalised by @xmath441 .
we define the pairing @xmath442 by @xmath443 straightforward computation shows that for @xmath444 we have @xmath445 which means for @xmath446 $ ] .
\(ii ) follows from the construction .
\(iii ) is proved in the same way as proposition 3.26 of @xcite .
let @xmath447 be a basis of @xmath448 and @xmath449 be its dual basis of @xmath450 with respect to @xmath442 .
[ prop : hat - i ] there exists an isomorphism @xmath451 , \label{hat - i}\ ] ] of @xmath373-modules defined by @xmath452 ) : = \bigoplus_{\lambda } \left[\sum_{d=0}^\infty \sum_{i } e_{\lambda , d , i } { \otimes}v { \otimes}e_{\lambda , d}^i \right ] q^d , \label{def : hat - i}\ ] ] for @xmath411 .
first we prove the well - definedness of , for which it is enough to show the well - definedness of its component @xmath453 $ ] , namely , @xmath454 for @xmath410 and @xmath455 .
since @xmath456 we have only to show that the left hand side of is equal to the left hand side of , which is equivalent to an equation in @xmath457 : @xmath458 recall that the germ of @xmath381 at @xmath233 and the germ at @xmath275 is related by @xmath459 .
see of this paper or ( 18 ) of @xcite . hence using the expansion and the invariance of the pairing
, we can show in the same way as the proof of claim 3 in the proof of theorem 6.2.1 in @xcite .
thus @xmath460 is well - defined .
obviously the @xmath94 part of @xmath460 is the isomorphism @xmath461 , .
therefore by termwise approximation ( in analytic language ) or , in other words , by nakayama s lemma ( in algebraic language ) , @xmath460 is shown to be an isomorphism . with these preparations ,
the proof of the locally freeness of @xmath196 goes as follows . as is mentioned after the statement of theorem [ thm : loc - free ] ,
it is enough to prove that the stalk @xmath367 at @xmath371 is a free @xmath369-module . since @xmath462 is isomorphic to @xmath463 ( lemma [ lem : completion ] ) and to @xmath464 $ ] ( proposition [ prop : hat - i ] )
, we have an isomorphism @xmath465 .
\label{cc = sum - ccorb : completed}\ ] ] the right hand side of being a free @xmath466-module ( proposition [ prop : loc - free : trig / orb ] ( ii ) and proposition [ prop : ccorb : integrable ] ) , faithfully flatness of @xmath466 over @xmath369 implies that @xmath367 is a free @xmath369-module .
thus theorem [ thm : loc - free ] is proved .
we have proved locally freeness of @xmath196 in two cases ; weyl module case ( proposition [ prop : coherence ] ) and integrable highest weight module case ( theorem [ thm : loc - free ] ) .
a few comments are in order : * in the weyl module case , @xmath467 as shown in the proof of proposition [ prop : coherence ] and the rank of @xmath196 is @xmath468 .
* in the integrable highest module case , the rank is computed by further degenerating the orbifold .
degeneration of the type @xmath111 should be considered in the same way as in @xcite or @xcite .
the final results of the degeneration is a combination of the three - punctured orbifold @xmath224 . in principle
a verlinde - type formula would be obtained in this way . * in @xcite we have shown that @xmath196 has a flat connection .
it has a regular singularity along @xmath469 , which is easily deduced from the explicit form of the connection , theorem 5.9 of @xcite .
hence there is a one - to - one correspondence between flat sections around @xmath212 and its restriction to @xmath212 or , in other words , the `` initial value '' at @xmath212 because of the locally flatness .
the author expresses his gratitude to akihiro tsuchiya who explained details of @xcite and @xcite , toshiro kuwabara who showed the manuscript of @xcite ( the best guide to @xcite for @xmath278 case ) before publishing , michio jimbo , gen kuroki , tetsuji miwa , hiroyuki ochiai , kiyoshi ohba , nobuyoshi takahashi , tomohide terasoma and yasuhiko yamada for discussion and comments .
the atmosphere and environment of institute for theoretical and experimental physics ( moscow , russia ) and the conference `` infinite dimensional algebras and integrable systems '' ( faro , portugal ) were very important .
the author thanks their hospitality .
tuy a. a. belavin , v. g. drinfeld , solutions of the classical yang - baxter equations for simple lie algebras .
i ego prilozh .
_ * 16 - 3 * , 129 ( 1982 ) ( in russian ) ; _ funct
* 16 * , 159180 ( 1982 ) ( english transl . )
y. shimizu , k. ueno , _ moduli theory iii _ , ( iwanami , tokyo , 1999 ) gendai suugaku no tenkai series ( in japanese ) ; _ advances in moduli theory _ , translations of mathematical monographs , * 206 * , iwanami series in modern mathematics , american mathematical society , providence , u.s.a .
( 2002 ) ( english translation ) t. takebe , trigonometric degeneration and orbifold wess - zumino - witten model .
i in _ the proceedings of the 6th international workshop on coformal and integrable models , chernogolovka , sep .
2002 _ , _ international journal of modern physics , a _ , * 19 * , supplement , 418435 ( 2004 ) a. tsuchiya , k. ueno , y. yamada , conformal field theory on universal family of stable curves with gauge symmetries . in
_ integrable systems in quantum field theory and statistical mechanics , adv . stud .
pure math .
_ * 19 * , 459566 ( 1989 ) .
k. ueno , on conformal field theory , in _ vector bundles in algebraic geometry ( durham , 1993 ) _ , ed . by n.
j. hitchin , p.
e. newstead and w. m. oxbury , _ london math .
lecture note ser . _ * 208 * , ( cambridge univ . press ,
cambridge , 1995 ) pp .
283345 , m. wakimoto , _ infinite - dimensional lie algebras _ , ( iwanami , tokyo , 1999 ) gendai suugaku no tenkai series ( in japanese ) ; translations of mathematical monographs , * 195 * , iwanami series in modern mathematics , american mathematical society , providence , u.s.a .
( 2001 ) ( english translation by k. iohara ) | the sheaves of conformal blocks and conformal coinvariants of the twisted wzw model have a factorisation property and are locally free even at the boundary of the moduli space , where the elliptic kz equations and the baxter - belavin elliptic @xmath0 matrix degenerate to the trigonometric kz equations and the trigonometric @xmath0 matrix , respectively .
etingof s construction of the elliptic kz equations is geometrically interpreted . |
the problem we consider here is to compute the exponential of an upper block - triangular , block - toeplitz matrix , that is , a matrix of the kind @xmath0,\ ] ] where @xmath1 , @xmath2 , are @xmath3 matrices .
our interest stems from the analysis in dendievel and latouche @xcite of the erlangization method for markovian fluid models , but the story goes further back in time .
the erlangian approximation method was introduced in asmussen _
@xcite in the context of risk processes ; it was picked up in stanford _
@xcite where a connection is established with fluid queues .
other relevant references are stanford _
@xcite where the focus is on modelling the spread of forest fires , and ramaswami _ et al . _
@xcite where some basic algorithms are developed .
markovian fluid models are two - dimensional processes @xmath4 where @xmath5 is a markov process with infinitesimal generator @xmath6 on the state space @xmath7 ; to each state @xmath8 is associated a rate of growth @xmath9 and @xmath10 is controlled by @xmath11 through the equation @xmath12 performance measures of interest include the distributions of @xmath10 and of various first passage times .
usually , @xmath11 is called the phase of the process at time @xmath13 and @xmath10 its level , and the phase space @xmath14 is partitioned into three subsets @xmath15 , @xmath16 and @xmath17 such that @xmath18 , @xmath19 or @xmath20 if @xmath8 is in @xmath15 , @xmath16 or @xmath17 , respectively . to simplify our presentation without missing any important feature , we assume below that @xmath17 is empty . the first return probabilities of @xmath10 to its initial level @xmath21 play a central role in the analysis of fluid queues .
it is customary to define two matrices @xmath22 and @xmath23 of first return probabilities : @xmath24 , \qquad \mbox{$i \in { \mathcal{s}}_+$ , $ j \in { \mathcal{s}}_-$,}\ ] ] and @xmath25 , \qquad \mbox{$i \in { \mathcal{s}}_-$ , $ j \in { \mathcal{s}}_+$,}\ ] ] where @xmath26 is the first passage time to level 0 .
thus , the entries of @xmath22 and @xmath27 are the probability of returning to the initial level after having started in the upward , and the downward directions , respectively .
if the process starts from some level @xmath28 , then @xmath29 = ( \begin{bmatrix } i
\\ \psi \end{bmatrix } e^{hx})_{ij } \qquad \mbox{$i \in \{1 , \ldots , m\}$ , $ j \in { \mathcal{s}}_-$;}\ ] ] here , @xmath30 is a square matrix on @xmath31 and is given by @xmath32 where @xmath33 and @xmath34 are submatrices of the generator @xmath6 , indexed by @xmath31 and @xmath35 , respectively , and @xmath36 is a diagonal matrix with @xmath37 on the diagonal .
a similar equation holds for @xmath38 .
the matrices @xmath22 and @xmath23 are solutions of algebraic riccati equations and their resolution has been the object of much attention .
very efficient algorithms are available , and we refer to bini _ et al . _
@xcite and bean _ et al . _
@xcite .
the erlangian approximation method is introduced in @xcite to determine the detailed distribution of @xmath39 .
the idea is that , to compute the probability @xmath40 , \qquad \mbox{$x>0 $ , $ i \in \{1 , \ldots , m\}$}\ ] ] for a fixed value @xmath41 , it is convenient to replace @xmath41 by a random variable @xmath42 with an erlang distribution , with parameters @xmath43 for some positive integer @xmath44 .
the random variable @xmath42 has expectation @xmath41 and variance @xmath45 , so that @xmath46 is a good approximation of @xmath47 if @xmath44 is large enough . from a computational point of view
, the advantage is that one replaces systems of integro - differential equations by linear equations .
the long and the short of it is that the original system is replaced by the process @xmath48 with a two - dimensional phase @xmath49 on the state space @xmath50 and with the generator @xmath51 where @xmath52 .
the physical interpretation is that the absorbing state 0 is entered at the random time @xmath42 , and the component @xmath53 of the phase marks the progress of time towards @xmath42 . some authors ( for instance @xcite ) report that good approximations may be obtained with small values of @xmath44 . because of the toeplitz - like structure of @xmath54 , the matrices @xmath22 and @xmath30 are both upper block - triangular block - toeplitz and it is interesting to use the toeplitz structure in order to reduce the cost when @xmath44 is large .
this is done in @xcite for the matrix @xmath22 . here
we address the question of efficiently computing the exponential matrix @xmath55 for a given value of @xmath56 , where @xmath30 has the structure of .
we shall assume without loss of generality that @xmath57 .
we recall that the exponential function can be extended to a matrix variable by defining @xmath58 for more details on the matrix exponential and more generally on matrix functions we refer the reader to higham @xcite .
the matrix @xmath59 defined in is of order @xmath60 and it may be huge , since a larger @xmath44 leads to a better erlangian approximation , while the size @xmath61 of the blocks is generally small . the matrix @xmath59 is a subgenerator , i.e. , it has negative diagonal entries , nonnegative off - diagonal entries , and the sum of the entries on each row is nonpositive . since block - triangular block - toeplitz matrices are closed under matrix multiplication , it follows from that the matrix exponential @xmath62 is also an upper block triangular , block - toeplitz matrix ; in particular , the diagonal blocks of @xmath62 coincide with @xmath63 .
moreover , it is known that the matrix @xmath62 is nonnegative and substochastic .
the problem of the computation of the exponential of a generator has been considered in xue and ye @xcite and by shao et al .
@xcite , where the authors propose component - wise accurate algorithms for the computation .
these algorithms are efficient for matrices of small size . for the erlangian approximation problem ,
these algorithms are useless for the large size of the matrices involved .
recently , some attention has been given to the computation of the exponential of general toeplitz matrices by using arnoldi method ( lee _ et al . _
@xcite , pang and sun @xcite ) . in our framework ,
toeplitz matrices are block - triangular so that they form a matrix algebra .
this property is particularly effective for the design of efficient algorithms and we propose some numerical methods that exploit the block - triangular block - toeplitz structure and the generator properties . unlike the general methods ,
our algorithms allow one to deal with matrices @xmath59 of very large size .
two methods rely on spectral and computational properties of block - circulant and block @xmath64-circulant matrices ( bini @xcite , bini _ et al . _
@xcite ) and on the use of fast fourier transforms ( fft ) .
recall that block @xmath64-circulant matrices have the form @xmath65,\ ] ] and that a block - circulant matrix is a block @xmath64-circulant matrix with @xmath66 . for simplicity ,
we denote by @xmath67 the block @xmath68-circulant matrix @xmath69 .
since block @xmath64-circulant matrices can be block - diagonalized by fft @xcite , the computation of the exponential of an @xmath70 block @xmath64-circulant matrix with @xmath3 blocks can be reduced to the computation of @xmath44 exponentials of @xmath3 matrices .
these latter exponentials are independent from each other and can be computed simultaneously with a multi - core architecture at the cost of a single exponential .
the idea of the first method is to approximate @xmath62 by @xmath71 where @xmath72 and @xmath73 is sufficiently small .
we analyse the error and are thereby able to choose the value of @xmath64 which gives a good balance between the roundoff error and the approximation error .
in fact , the approximation error grows as @xmath74 while the roundoff error is @xmath75 , where @xmath76 is the machine precision .
this leads to an overall error which is @xmath77 . by using the fact that the solution is real , by choosing @xmath64 a pure imaginary number we get an approximation error @xmath78 which leads to an overall error @xmath79 .
since the approximation error is a power series in @xmath64 , we devise a further technique which consists in averaging the solutions computed with @xmath80 different values of @xmath64 .
this way , we are able to cancel out the components of the error of degree less than @xmath81 .
this leads to a substantial improvement of the precision .
moreover , since the different computations are independent from each other , the computational cost in a multicore architecture is independent of @xmath80 . in our second approach , the matrix @xmath59 is embedded into a @xmath82 block - circulant matrix @xmath83 , where @xmath84 is sufficiently large , and an approximation of @xmath85 is obtained from a suitable submatrix of @xmath86 .
the computation of @xmath86 is reduced to the computation of @xmath84 exponentials of @xmath3 matrices , and our error analysis allows one to choose the value of @xmath84 so as to guarantee a given error bound in the computed approximation .
the third numerical method consists in specializing the shifting and taylor series method of @xcite .
the block - triangular toeplitz structure is exploited in the fft - based matrix multiplications involved in the algorithm , leading to a reduction of the computational cost .
the algorithm obtained in this case does not seem well suited for an implementation in a multicore architecture .
we compare the three numerical methods , from a theoretical as well as from a numerical point of view . from our analysis , we conclude that the method based on @xmath64-circulant matrices is the fastest and provides a reasonable approximation to the solution .
moreover , by applying the averaging technique we can dramatically improve the accuracy .
the method based on embedding and the one based on power series perform an accurate computation but are slightly more expensive .
it must be emphasised that the use of fft makes the algorithms norm - wise stable but that component - wise stability is not guaranteed . in consequence , the matrix elements with values of modulus below the machine precision may not be well approximated in terms of relative error .
the paper is organised as follows . in sections
[ sec : der ] and [ sec : fastcomp ] , we recall properties of the exponential of a subgenerator and of its derivatives , and some basic properties of block - toeplitz and block - circulant matrices which are used in our algorithms . in section [ sec : expcirc ] , we show how to compute the exponential of a block @xmath64-circulant matrix by using fast arithmetic based on fft and we perform an error analysis .
we present in section [ sec : exptt ] the algorithms to compute the exponential of @xmath59 : first we analyse the decay of off - diagonal entries of the matrix exponential , next we describe the new methods and perform an error analysis .
we conclude with numerical experiments in section [ sec : exper ] .
a subgenerator of a markov process is a matrix @xmath54 of real numbers such that the off diagonal entries of @xmath54 are nonnegative , the diagonal entries are negative , and the sum of the entries on each row is nonpositive .
we denote by @xmath87 the column vector with all entries equal to 1 , with size according to the context .
if @xmath54 is a subgenerator , then @xmath88 and @xmath54 is called a generator if the row sum on all rows is zero .
let @xmath89 .
the matrix @xmath90 is a nonnegative matrix , and we may write @xmath91 . from the latter equality it follows that the matrix exponential @xmath92 is nonnegative
. moreover , since @xmath88 it follows that @xmath93 .
therefore , in view of , @xmath94 .
thus we may conclude that @xmath95 , that is , @xmath92 is a substochastic matrix .
we recall the definition and some properties of the gteaux and frchet derivatives , and their expression for the matrix exponential function , together with some properties when the matrix is a subgenerator .
we refer the reader to @xcite for more details .
recall that if @xmath111 is a subgenerator , then @xmath108 is a substochastic matrix for any @xmath112 and in particular @xmath113 .
therefore , by taking norms in , we obtain the upper bound @xmath114 which may be extended to the @xmath107-th order gteaux derivative as in the next proposition .
[ prop1 ] if @xmath111 is a subgenerator , then @xmath115}(tx , e)\|_\infty \le t^j \| e\|_\infty^j\ ] ] for @xmath116 , for any @xmath117 . moreover , if @xmath100 is a nonnegative matrix , then @xmath118}(tx , e)$ ] is nonnegative for any @xmath117 .
since the matrix @xmath111 is a subgenerator , the matrix @xmath119 is nonnegative and substochastic for any @xmath120 , therefore @xmath121 for any @xmath120 . by using this property ,
the inequality can be proved by induction . if @xmath122 , then @xmath123}(tx , e)\|_\infty \le 1 $ ]
. the inductive step is immediately proved , since from we have @xmath124}(tx , e)\|_\infty & \le j\int_0^t \|e^{(t - s)x}\|_\infty \|e\|_\infty \|g^{[j-1]}(sx , e)\|_\infty ds\\ & \le j \int_0^t \|e\|_\infty^j s^{j-1 } ds=\|e\|_\infty^j t^j , \end{split}\ ] ] where the last inequality follows from the inductive assumption . if the matrix @xmath100 is nonnegative , from the recurrence and from the fact that @xmath119 is nonnegative and substochastic for any @xmath120 , it follows by induction that @xmath106}(tx , e)$ ] is nonnegative for any @xmath117 .
@xmath125 the following result provides some bounds related to the exponential of the matrix @xmath59 of and to its gteaux derivative ; it will be used in the next sections to analyse the stability of the algorithm in section [ sec : epsc ] based on @xmath64-circulant matrices .
[ thm : new ] let @xmath126 be the matrix in and assume it is a subgenerator . for @xmath127 define @xmath128 .
if @xmath129 , @xmath130 , then @xmath131 for any @xmath132 . moreover , @xmath133 for any @xmath3 matrix @xmath100 . proof .
define @xmath134 , @xmath135 and @xmath136 . from the choice of @xmath137 it follows that @xmath138 .
we have @xmath139 where the latter inequality holds in view of ( * ? ? ?
* theorem 10.10 ) . since @xmath138 we may write that @xmath140 . from the inequality @xmath141 we find that @xmath142 , that is @xmath143 whence @xmath144 .
therefore we conclude that @xmath145 and the first claim is proved .
+ now , concerning @xmath146 we have @xmath147 since @xmath148 , @xmath149 and @xmath150 , @xmath151 , we have @xmath152 by monotonicity of the infinity norm , we have @xmath153 , @xmath154 and , from , @xmath155 from , we have @xmath156 and the last claim follows . @xmath125
in this section we recall some basic properties of block - toeplitz and block - circulant matrices , useful for our computational analysis .
we refer the reader to bini and pan @xcite and bini _ et al . _
@xcite for more details . given a matrix @xmath157 , we denote by @xmath158 and by @xmath159 the transpose matrix and the transpose conjugate matrix of @xmath160 , respectively .
the conjugate of a complex number @xmath161 is denoted by @xmath162 .
let @xmath163 be the imaginary unit such that @xmath164 and @xmath165 be a primitive @xmath44th root of the unity .
we denote by @xmath166 the fourier matrix .
recall that @xmath167 is nonsingular , that @xmath168 and that , given a vector @xmath169 , the application @xmath170 defines the inverse discrete fourier transform ( idft ) of @xmath171 .
we assume that @xmath44 is an integer power of 2 , so that the vector @xmath172 can be computed by means of the fft algorithm in @xmath173 arithmetic operations ( ops ) .
the application @xmath174 is called discrete fourier transform ( dft ) and the vector @xmath171 can be computed in @xmath175 ops .
given the @xmath3 matrices @xmath176 , @xmath2 , we denote by @xmath177 the block-(column ) vector with block - entries @xmath176 , @xmath2 .
finally , we define @xmath178 , where @xmath179 is the kronecker product and @xmath180 the identity matrix of order @xmath61 . this way , for a block - column vector @xmath160 the matrix @xmath181 can be computed by means of @xmath182 idfts with @xmath183 ops .
similarly , given the matrix @xmath184 , the block - vector @xmath185 can be computed with @xmath186 ops .
for the results in this section we refer the reader to the book @xcite and to the references cited therein .
given the block - vector @xmath187 , with @xmath3 blocks , the @xmath188 block - circulant matrix @xmath189 associated with @xmath184 is the matrix with block - entries @xmath190 so that @xmath191 $ ] coincides with the first block - row of @xmath192 and the entries of any other block - row are obtained by the entries of the previous block - row by a cyclic permutation which moves the last block entry to the first position and shifts the remaining block - entries one place to the right . for instance , for @xmath193 one has @xmath194.\ ] ] observe that a block - circulant matrix is a particular block - toeplitz matrix .
block - circulant matrices can be simultaneously block - diagonalized by means of fft , that is , @xmath195 this property shows that block - circulant matrices are closed under matrix multiplication , i.e. , they form a matrix algebra , moreover the product of a circulant matrix and a vector can be computed by means of algorithm [ alg:1 ] .
this algorithm performs the computation with @xmath196 ffts and @xmath44 matrix multiplications . since @xmath197
ops are sufficient to multiply two @xmath3 matrices , the overall cost of algorithm [ alg:1 ] is @xmath198 ops . @xmath199 @xmath200 @xmath201 , @xmath2 , @xmath202 @xmath203 if the input block - vectors are real then the vectors @xmath199 and @xmath204 have a special structure , that is , the components @xmath205 and @xmath206 are real while @xmath207 , @xmath208 , for @xmath209 . in this case , the number of matrix multiplications at step 3 of algorithm [ alg:1 ] is reduced to @xmath210 .
[ rem : cir]observe that the product of two circulant matrices may be computed by means of a product of a circulant matrix and a vector by means of algorithm [ alg:1 ] .
in fact , since the last column of the block - circulant matrix @xmath192 is the block - vector @xmath211 , if @xmath212 then we find that @xmath213 , where @xmath214 , @xmath215 .
we denote by @xmath184 the block - vector @xmath216 and by @xmath59 the block - upper triangular block - toeplitz matrix whose first row is @xmath191 $ ] . for @xmath193 , for instance , @xmath217.\ ] ] block - upper triangular block - toeplitz matrices are closed under matrix multiplication .
consider the vector @xmath218 of @xmath219 components obtained by filling the vector @xmath184 with zero blocks , and the block - vector @xmath177 such that @xmath220 and @xmath221 for @xmath130 .
then the matrix @xmath222 can be partitioned as follows @xmath223,\ ] ] where @xmath224 is the block - lower triangular block - toeplitz matrix whose first block - column is @xmath160 .
this expression enables one to compute the product @xmath225 of a block - upper triangular toeplitz matrix and a block - vector with a low number of arithmetic operations .
in fact , from one deduces that @xmath226 coincides with the first half of the block - vector @xmath227 where @xmath228 is the block - vector of length @xmath219 obtained by filling @xmath111 with zeros .
this fact leads to algorithm [ alg:2 ] for computing the product of a block - triangular block - toeplitz matrix and a block - vector .
the cost of this algorithm is @xmath229 ops .
set @xmath230 with @xmath231 for @xmath2 , @xmath232 for @xmath233 set @xmath234 with @xmath235 for @xmath2 , @xmath236 for @xmath233 apply algorithm [ alg:1 ] to compute @xmath227 set @xmath237 with @xmath238 , for @xmath2 .
[ rem : tt]observe that the product of two block - triangular block - toeplitz matrices can be computed by means of a product of a block - triangular block - toeplitz matrix and a block - vector by means of algorithm [ alg:2 ] .
in fact , since the last column of @xmath59 is the block vector @xmath211 , if @xmath239 then we find that @xmath240 , where @xmath214 , @xmath215 . given a block - vector @xmath184 and a complex number @xmath64 ,
the block-@xmath64-circulant matrix @xmath241 is defined by @xmath242 for instance , for @xmath193 one has @xmath243.\ ] ] observe that a block-@xmath64-circulant matrix is a particular case of block - toeplitz matrix and that , for @xmath73 small , a block @xmath64-circulant matrix is an approximation of a block - triangular block - toeplitz matrix . like block - circulant matrices
, block-@xmath64-circulant matrices can be simultaneously block - diagonalized by means of fft , so that they are closed under matrix multiplication and form a matrix algebra as well .
in fact , one can show that @xmath244 where @xmath245
let @xmath187 be a block - vector of length @xmath44 where @xmath246 , consider the block-@xmath64-circulant matrix @xmath247 and its matrix exponential @xmath71 . in view of
, we find that @xmath248 therefore the exponential of a block-@xmath64-circulant matrix is still block-@xmath64-circulant .
moreover , we have @xmath249 where @xmath250 and @xmath251 .
the above equations allow to compute the exponential of an @xmath188 block-@xmath64-circulant matrix by computing @xmath44 exponentials of @xmath3 matrices and two fourier transforms , as described in algorithm [ alg : expepsc ] .
@xmath252 @xmath253 @xmath254 , @xmath2 , and set @xmath202 @xmath255 @xmath256 observe that the multiplication of @xmath184 by the diagonal matrix @xmath257 at step 1 reduces to scaling the blocks @xmath1 by the scalar @xmath258 .
the multiplication by @xmath259 at step 5 performs similarly .
therefore the overall cost of the algorithm is given by @xmath260 ops plus the cost of computing @xmath44 exponentials of @xmath3 matrices .
for @xmath66 the block-@xmath64-circulant matrix turns to a block - circulant matrix and algorithm [ alg : expepsc ] takes the simpler form described in algorithm [ alg : expcirc ] . the computational cost in this case
is reduced to @xmath261 ops plus the cost of computing @xmath44 exponentials of @xmath3 matrices .
@xmath262 @xmath254 , @xmath2 , and set @xmath202 @xmath263 let @xmath187 be the block - vector defining the first block row of the subgenerator @xmath59 .
we analyze the error generated by computing the exponential of the block-@xmath64-circulant matrix @xmath247 by means of algorithm [ alg : expepsc ] in floating point arithmetic , where @xmath264 with @xmath265 . here and hereafter
@xmath266 denotes the result computed in floating point arithmetic of the expression between parenthesis .
the symbol @xmath267 denotes equality up to lower order terms , and similarly the symbol @xmath268\leq\end{array}}$ ] stands for inequality up to lower order terms .
the symbol @xmath76 denotes the machine precision .
we recall the following useful fact ( see @xcite ) @xmath269 for @xmath270 , and we use the following properties involving norms , where @xmath169 @xmath271 in order to perform the error analysis of algorithm [ alg : expepsc ] , we recall the following result concerning fft ( see @xcite ) .
[ th : fft ] let @xmath56 be a vector of @xmath44 components , @xmath272 , @xmath273 integer , @xmath274 , where @xmath166 is the fourier matrix .
let @xmath275 be the vector obtained in place of @xmath276 by applying the cooley - tukey fft algorithm in floating point arithmetic with precision @xmath76 where the roots of the unity are approximated by floating point numbers up to the error @xmath277 .
then @xmath278 in particular , with @xmath279 and performing a first - order error analysis where we consider only the part of the error which is linear in @xmath76 we have @xmath280 observe that , since @xmath281 , we may replace @xmath167 with @xmath282 in the statement of theorem [ th : fft ] .
we split algorithm [ alg : expepsc ] into three parts .
the first part consists in computing the entries of the matrices @xmath283 by means of steps 1 and 2 , the second part consists in computing the entries of @xmath284 and the third part is formed by the remaining steps 4 and 5 .
the first and third part can be viewed as the collection of @xmath182 independent computations applied to the entry @xmath285 of the generic block for @xmath286 .
more specifically , given the pair @xmath285 , denote @xmath287 , @xmath288 , @xmath289 the vectors whose components are @xmath290 , @xmath291 , @xmath292 , @xmath293 , respectively .
the computation of @xmath172 is obtained in the following way : @xmath294 , @xmath295 , for @xmath293 , @xmath296 . while , denoting @xmath297 the vectors whose components are @xmath298 , @xmath299 , @xmath300 , @xmath293 , respectively , the computation of @xmath276 is obtained in the following way : @xmath301 , @xmath302 for @xmath293 .
define @xmath303 , @xmath304 , @xmath305 , @xmath306 where @xmath307 are the values obtained in place of @xmath308 by performing computations in floating point arithmetic .
we denote also by @xmath309 and @xmath310 the @xmath80-th component of @xmath311 and @xmath312 , respectively .
in our analysis we assume that the constants @xmath313 have been precomputed and approximated with the numbers @xmath314 such that @xmath315 , @xmath316 , @xmath317 . since @xmath318 , from we find that @xmath319 .
thus , @xmath320 denoting by @xmath321 the error introduced in computing the fft @xmath171 of @xmath161 in floating point arithmetic , we have @xmath322 and in view of , , and we obtain @xmath323 where the last inequality follows from the fact that @xmath324 since @xmath295 and @xmath325 .
this implies that @xmath326 is such that @xmath327 which yields @xmath328 concerning the second part of the computation , for the matrix @xmath329 we have @xmath330 where @xmath331 is the error generated by computing the matrix exponential in floating point arithmetic .
here we assume that @xmath332 for some positive constant @xmath39 which depends on the algorithm used to compute the matrix exponential . from the properties of the gteaux derivative one
has @xmath333 , and from theorem [ thm : new ] , applied with @xmath334 , @xmath130 , it follows that @xmath335 and @xmath336 . combining these results with leads to the bound @xmath337\leq\end{array}}\|\delta_{v_k}\|_\infty+\mu\tau.\ ] ] finally , for the third part of the computation , consisting of steps 4 and 5
, we have @xmath338 where @xmath339 is the error obtained by computing @xmath340 in floating point arithmetic .
thus from we have @xmath341\leq\end{array}}\frac 1n \|f\delta_w\|_\infty+ \mu \gamma \log_2n\|r\|_2\\ &
\le \|\delta_w\|_\infty+\mu\gamma \sqrt n\log_2n\|y\|_\infty , \end{split}\ ] ] where the second inequality holds from and from @xmath342 since @xmath343 . moreover , we find that @xmath344 now we are ready to combine all the pieces and obtain the error bound on the computed value @xmath226 . from
we get @xmath345\leq\end{array}}|\theta|^{-k}(\|\delta_w\|_\infty + ( \zeta+\gamma \sqrt n\log_2n)\mu\|y\|_\infty).\ ] ] thus we have @xmath346\leq\end{array}}m|\theta|^{-k}(\max_h\|\delta_{w_h}\|_\infty+ ( \zeta+\gamma \sqrt n\log_2n)\mu \max_h\|y_h\|_\infty).\ ] ] moreover , from and we conclude with the following bound @xmath347\leq\end{array}}m|\theta|^{-k } ( \max_h\|\delta_{v_h}\|_\infty+\mu\tau + ( \zeta+\gamma \sqrt n\log_2n)\mu \max_h\|y_h\|_\infty).\ ] ] whence @xmath347\leq\end{array}}\mu m|\theta|^{-k } ( m n(\zeta+\gamma \log_2n)\max_{r , s , h } \max_h\|y_h\|_\infty)\ ] ] and we may conclude with the following [ th : err ] let @xmath348 be the value of @xmath349 provided by algorithm [ alg : expepsc ] applied in floating point arithmetic with precision @xmath76 for computing @xmath350 , where @xmath351 , @xmath352 .
denote @xmath353 .
one has @xmath347\leq\end{array}}\mu\epsilon^{-1 } m\varphi\ ] ] where @xmath354 @xmath355 , @xmath356 , and @xmath357 is the error bound in the computation of the matrix exponential , i.e. , such that @xmath358 for an @xmath3 matrix @xmath160 . in the case where @xmath66 , we apply algorithm [ alg : expcirc ] to compute the exponential of a block - circulant matrix and the above result leads to [ th : err2 ] let @xmath348 be the value of @xmath349 provided by algorithm [ alg : expcirc ] applied in floating point arithmetic with precision @xmath76 for computing @xmath359 , where @xmath351 , @xmath352
. denote @xmath353 .
one has @xmath360 where latexmath:[$\chi = m n
\gamma \log_2n \max_{r , s , h } @xmath355 , @xmath356 , and @xmath357 is the error bound in the computation of the matrix exponential , i.e. , such that @xmath358 for an @xmath3 matrix @xmath160 .
let @xmath187 be the block - vector defining the first block - row of the subgenerator @xmath59 of .
since block - triangular block - toeplitz matrices form a matrix algebra , by using the taylor series expansion of the matrix exponential , it follows that @xmath62 is still a block - triangular block - toeplitz matrix .
denote by @xmath362 the block - vector defining the entries on the first block - row of @xmath62 , i.e. , such that @xmath363 .
in particular , we have @xmath364 .
let @xmath365 and define the @xmath84 dimensional block - vector @xmath366 obtained by completing @xmath184 with zeros : @xmath367 consider the @xmath82 block - triangular block - toeplitz matrix @xmath368 .
in view of ( * ? ? ? * theorem 3.6 ) , if @xmath369 , then @xmath370 is the principal @xmath371 block - submatrix of @xmath372 .
denote by @xmath373 the block - vector defining the first block - row of @xmath374 , i.e. , @xmath375 is the block - vector such that @xmath376 .
let @xmath377 be such that @xmath378 where @xmath379 .
define the block - vector @xmath380 with block - components @xmath381 for @xmath2 , and @xmath382 for @xmath383 . observe that @xmath384 is a nonnegative matrix , and we may write @xmath385 .
we denote by @xmath386 the block - vector such that @xmath387 .
in particular we have @xmath388 , @xmath389 . in this section
we investigate decay properties of the exponential @xmath374 of a subgenerator , in the case where the subgenerator is a banded block - triangular block - toeplitz matrix .
these properties will be used in section [ sec : emb ] to estimate the approximation error of the numerical method based on the embedding into a block - circulant matrix .
decay properties of matrix functions have been analyzed in the literature .
we refer the reader to the survey paper @xcite and to @xcite . in our case
the structure and sign properties play an important role .
the matrix exponential @xmath374 is not banded in general , but its off - diagonal entries have useful decay properties for @xmath390 . to prove this fact we need the following result ( * ? ? ?
* theorem 3.6 ) on decay properties of analytic functions .
[ thm : dec ] let @xmath391 be an @xmath3 matrix power series analytic for @xmath392 with @xmath393 , and @xmath394 . for any @xmath395 , the block - coefficients satisfy @xmath396 is the @xmath3 matrix with elements @xmath397 , for @xmath286 , and the inequality is meant componentwise .
the following result provides bounds to @xmath398 , @xmath389 .
[ thm : decexp ] let @xmath365 and let @xmath399 , with @xmath400 , where @xmath126 in is a subgenerator and @xmath366 is defined in . for any @xmath401 , we have @xmath402 where @xmath379 .
we associate with the block - vector @xmath377 of the @xmath3 matrix polynomial @xmath403 .
for the properties of block triangular block - toeplitz matrices @xcite , the matrix @xmath404 is still a block - triangular block - toeplitz matrix and the blocks in its first row are the coefficients of the matrix polynomial @xmath405 .
let @xmath406 be the matrix coefficient of degree @xmath8 of @xmath407 , for @xmath389 .
from the power series expression of the matrix exponential we find that @xmath408 we want to give an upper bound to the matrices @xmath409 . since @xmath410 is a matrix polynomial , then it is analytic in all the complex plane and we may apply theorem [ thm : dec ] with @xmath411 and any @xmath401 .
we have to estimate the matrix @xmath412 .
the matrix coefficients of @xmath413 are nonnegative , therefore for any @xmath401 and for any @xmath392 with @xmath414 , we have @xmath415 . since @xmath126 is a subgenerator then @xmath416 .
so that we obtain @xmath417 .
since @xmath405 , then @xmath406 is the matrix coefficient of degree @xmath8 of @xmath410 and , in view of , we find that @xmath418 . from this inequality and from
we obtain that for any @xmath401 and for @xmath389 @xmath419 since @xmath420 we conclude the proof .
@xmath125 let @xmath72 with @xmath73 sufficiently small , and consider the block-@xmath64-circulant matrix @xmath421.\ ] ] the exponential of @xmath247 is still a block-@xmath64-circulant matrix , that can be computed by means of algorithm [ alg : expepsc ] .
denote by @xmath237 the block - vector such that @xmath422 .
the idea is to approximate the blocks @xmath398 , defining @xmath363 , by the matrices @xmath423 , for @xmath2 . in order to estimate the approximation error ,
observe that the matrix @xmath247 can be written as @xmath424 , where @xmath425.\ ] ] this property allows to give the following estimate : [ thm1 ] assume that @xmath59 is a subgenerator , and that @xmath264 .
one has @xmath426 moreover , if @xmath64 is a pure imaginary number , then @xmath427 where @xmath428 is the real part of @xmath71 .
proof . according to , @xmath429}({\mathcal{t}}(u),l),\ ] ]
where @xmath106}({\mathcal{t}}(u),l)$ ] are defined by means of . from proposition [ prop1 ]
we obtain @xmath430}({\mathcal{t}}(u),l)\|_\infty \le \sum_{j=1}^\infty \frac{|\epsilon|^j}{j ! } \|l\|_\infty^j= e^{|\epsilon|\|l\|_\infty}-1.\ ] ] if @xmath64 is a pure imaginary number , since @xmath62 is a real matrix , the inequality is obtained by comparing the real parts in and by applying proposition [ prop1 ] . @xmath125 it is interesting to observe that the choice of an imaginary value for @xmath64 provides an approximation error of the order @xmath431 instead of @xmath432 .
the idea of using an imaginary value for @xmath64 was used in @xcite in the framework of frchet derivative approximation of matrix functions .
the error bound can be improved by performing the computation with several different values of @xmath64 and taking the mean of the real parts of the results obtained this way .
for instance , choose @xmath433 , @xmath434 , where @xmath435 , and recall that @xmath436 , @xmath437 are power series in @xmath64 . taking the arithmetic mean of @xmath438 and @xmath439 , the components of odd degree in @xmath64 cancel out while the coefficient of @xmath440 is pure imaginary . therefore taking the real part of the arithmetic mean
provides an error @xmath441 .
this technique can be generalized as follows .
choose an integer @xmath442 and set @xmath443 , @xmath444 , where @xmath445 is a principal @xmath80-th root of @xmath163 .
then one can verify that the arithmetic mean of @xmath436 , @xmath444 is a power series in @xmath446 , moreover , @xmath447 is a pure imaginary number so that the real part of this mean provides an approximation with error @xmath448 .
observe that computing the exponential for different values of @xmath64 might seem a substantial computational overload .
however , in a parallel model of computation , the exponentials @xmath449 , @xmath444 , can be computed simultaneously by different processors at the same cost of computing a single exponential .
algorithm [ alg : epscirc_int ] reports this averaging technique .
set @xmath443 , @xmath444 compute the first block row @xmath450 of @xmath451 , @xmath444 , by means of algorithm [ alg : expepsc ] . set @xmath452 theorem [ thm1 ] provides us with a bound on the error generated by approximating the exponential of a block - upper triangular toeplitz matrix by means of the exponential of a block-@xmath64-circulant matrix .
in fact , in practical computations in floating point arithmetic , the overall error is formed by two components : one component is given by the approximation error analyzed in theorem [ thm1 ] , the second component is due to the roundoff and is estimated by theorem [ th : err ] .
more precisely , the effectively computed approximation in floating point arithmetic is the block - vector with components @xmath453 , @xmath293 , where @xmath454 is bounded in theorem [ th : err ] . on the other hand ,
@xmath455 where , by theorem [ thm1 ] , @xmath456 is such that @xmath457\|_\infty\le \left\{\begin{array}{ll } \psi(|\epsilon|^2\| l\|_\infty^2 ) & \hbox{if~ } \epsilon \hbox{~is imaginary}\\[1ex ] \psi(|\epsilon|\| l\|_\infty ) & \hbox{otherwise } \end{array}\right.\ ] ] where @xmath458 .
this way , for the overall error @xmath459 one has @xmath460 for @xmath461 , or @xmath462 .
this shows the need to find a proper balance between the two errors : small values for @xmath73 provide a small approximation error @xmath463 but the roundoff errors diverge to infinity as @xmath464 .
a good compromise is to choose @xmath64 so that the upper bounds to @xmath465 and @xmath454 have the same order of magnitude . equating these upper bounds in the case of non - imaginary @xmath64 yields @xmath466\leq\end{array}}2\epsilon \|l\|_\infty\ ] ] and in the case of imaginary @xmath64 , @xmath467{m\mu\varphi/\|l\|_\infty^2},\quad \|e_k\|_\infty{\begin{array}{c}^{\displaystyle \cdot}\\[-2ex]\leq\end{array}}2\epsilon^2\|l\|_\infty^2.\ ] ] the latter bound is an @xmath79 .
this implies that asymptotically , as @xmath468 , we may loose @xmath469 of the digits provided by the floating point arithmetic .
if we adopt the strategy of performing the computation with @xmath80 different values of @xmath443 , @xmath444 , so that the approximation error is @xmath448 , then the total error turns to @xmath470 , i.e. , only @xmath471 digits are lost .
an interesting point is that the quantities @xmath472 and @xmath473 are involved in the expressions of the error bound .
since @xmath59 is a generator , both these quantities are bounded from above by @xmath379 . however , by means of simple manipulations , we may scale the input so that it is bounded by 1 .
this is performed by applying to @xmath59 the scaling and squaring technique of @xcite .
let @xmath474 be an integer such that @xmath475 . then ,
since @xmath476 , we first compute @xmath477 and then recover @xmath62 by performing @xmath478 repeated matrix squaring . in this way we have @xmath479 and @xmath480 .
since @xmath481 is still a generator , the error analysis performed for @xmath62 applies as well , and we can approximate the first block - row of @xmath477 with the first block - row @xmath482 of @xmath483 for a suitable @xmath264 with @xmath265 . finally we recover an approximation to @xmath62 by computing @xmath484 by means of @xmath478 repeated squarings , by using the toeplitz structure and algorithm [ alg:2 ] , in view of remark [ rem : tt ] .
the overall procedure is described in algorithm [ alg : ttepsc ] .
@xmath379 , @xmath485 and @xmath486 compute @xmath226 , the first block - row of @xmath487 , by means of algorithm [ alg : expepsc ] if @xmath64 is imaginary , replace @xmath226 with the real part of @xmath226 the idea of this method is to embed the matrix @xmath126 into a @xmath488 block - circulant matrix @xmath489 .
the first block - row of @xmath490 is approximated by the first @xmath44 blocks of the first block - row of @xmath491 .
specifically , take @xmath365 and consider the block - vector @xmath366 defined in .
the block - circulant matrix @xmath489 may be partitioned as @xmath492,\ ] ] where @xmath493 and @xmath54 are @xmath494 and @xmath495 block - matrices , respectively .
denote by @xmath496 and by @xmath497 the @xmath498 matrices formed by the first @xmath499 and the last @xmath499 columns , respectively , of the identity matrix of size @xmath500 .
the matrix @xmath489 can be also written as @xmath501 where the matrix @xmath502 is defined in .
because of the triangular toeplitz structure , the desired matrix @xmath503 is identical to the @xmath188 block - leading submatrix of @xmath504 .
our idea is to approximate the first block - row of @xmath503 with the first @xmath44 blocks of the first row of @xmath491 . as pointed out in section [ sec : expcirc ] , @xmath491 is a block - circulant matrix , and can be computed by means of algorithm [ alg : expcirc ] with @xmath505 ops , plus the cost of computing @xmath84 exponentials of @xmath506 matrices .
denote by @xmath507 the first block - row of @xmath508 , so that @xmath509 . an approximation of the matrices @xmath398 , @xmath2 ,
defining the first block - row of @xmath510 is provided by @xmath511 , @xmath2 ; as @xmath84 increases , the approximation improves , as shown by the following result .
[ thm : erremb ] let @xmath512 , with @xmath362 .
let @xmath365 and let @xmath513 , with @xmath507 .
one has @xmath514 for @xmath2 , and @xmath515 \right\|_\infty \le f_k(\sigma)\ ] ] for any @xmath401 , where @xmath516 with @xmath379 and @xmath502 defined in . proof . by using and
, we find that @xmath517}(\mathcal t(u^{(k)}),h_k).\ ] ] equating the first @xmath44 blocks in the first block - row in the above equation yields @xmath518= \sum_{j=1}^\infty \frac{1}{j ! }
w^{[j]},\ ] ] where @xmath519}$ ] is the block - row vector formed by the first @xmath44 block - entries in the first block - row of @xmath118}(\mathcal t_k(u),h_k)$ ] .
that is , @xmath520}=\widehat{\mathcal e}_1^t g^{[j]}(\mathcal t(u^{(k)}),h_k)\mathcal e_1,\ ] ] where @xmath521 is the @xmath522 matrix formed by the first @xmath61 columns of the identity matrix .
since @xmath523 , from and from proposition [ prop1 ] we deduce that @xmath519}\ge 0 $ ] so that @xmath524 for @xmath2 . on the other hand , in view of and from the fact that @xmath525 , we may write @xmath526 } & = j \int_0 ^ 1 \widehat{\mathcal e}_1^t e^ { ( 1-s)\mathcal t(u^{(k ) } ) } \mathcal e_k l \mathcal e_1^t g^{[j-1]}(s\mathcal t(u^{(k)}),h_k ) \mathcal e_1 ds\\ & = j \int_0 ^ 1
v(s ) l z^{[j-1]}(s)ds \end{split}\ ] ] where @xmath527 $ ] is the block - row vector formed by the last @xmath44 block - entries of the first block - row of @xmath528 , and @xmath529}(s)$ ] is the @xmath188 block leading submatrix of @xmath530}(s { \mathcal{t}}(u^{(k)}),h_k)$ ] .
since the matrix @xmath531 is a subgenerator , it follows that @xmath532 and , from theorem [ thm : decexp ] , that for any @xmath401 , @xmath533 for @xmath2 , where the latter inequality follows from the fact that @xmath534 .
this implies that @xmath535 moreover , since @xmath536 is a subgenerator , @xmath537 is nonnegative and @xmath538 , then , from proposition [ prop1 ] , we have @xmath530}(s{\mathcal{t}}(u^{(k)}),h_k)\ge 0 $ ] and @xmath539}(s{\mathcal{t}}(u^{(k)}),h_k)\|_\infty\le s^{j-1}\| l\|_\infty^{j-1}.\ ] ] this latter inequality implies that @xmath540}(s)\|_\infty\le s^{j-1}\| l\|_\infty^{j-1}$ ] .
therefore , by taking norms in , we find that @xmath541}\|_\infty & \le j \int_0 ^ 1 \| v(s)\|_\infty \| l \|_\infty \| z^{[j-1]}(s)\|_\infty ds \\ & \le j \| l\|_\infty^{j } e^{\alpha(\sigma^{n-1}-1 ) } \frac{\sigma^{-k+n}}{1-\sigma^{-1 } } \int_0 ^ 1 s^{j-1 } ds = \| l\|_\infty^{j } e^{\alpha(\sigma^{n-1}-1)}\frac{\sigma^{-k+n}}{1-\sigma^{-1}}. \end{split}\ ] ] hence , by taking norms in , we obtain . @xmath125 the matrices
@xmath398 and @xmath511 have a probabilistic interpretation .
namely , the matrix @xmath398 is the probability that the bmap is absorbed after time 1 , _ and _ at time 1 there have been @xmath542 arrivals ; the matrix @xmath511 is the probability that the bmap is absorbed after time 1 , _ and _ at time 1 there have been @xmath8 , or @xmath543 , or @xmath544 , or , arrivals . clearly , there are more trajectories favourable for @xmath511 than for @xmath398 and @xmath545
. similarly , there are more trajectories favourable for @xmath511 than for @xmath546 for a positive integer @xmath547 .
this shows that , if we take a sequence of integers @xmath548 , @xmath549 , , and a sequence @xmath550 , @xmath551 , @xmath552 , , such that @xmath553 , then @xmath554 for @xmath555 . therefore , the sequence @xmath556 has some monotonicity property in its convergence to @xmath6 .
the bound in shows that the error has an exponential decay as @xmath84 increases . moreover , such bound holds for any @xmath557 .
therefore we can fix a tolerance @xmath64 and a @xmath401 , and find @xmath84 such that @xmath558 . since we would like to keep @xmath84 as low as possible ,
another way to proceed is to fix a tolerance @xmath64 and find @xmath559 such that the size @xmath84 for which @xmath558 is minimum .
more specifically , after some manipulations , from the condition @xmath558 we obtain that @xmath560 where @xmath561 since @xmath401 is arbitrary , we choose @xmath559 such that @xmath562 has a minimum value .
in fact , the function @xmath562 diverges to infinity as @xmath559 tends to 1 and to @xmath563 , therefore it has at least a local minimum @xmath564 and we can choose @xmath565 .
when we perform the computation in floating point arithmetic , we have to consider also the error generated by roundoff in computing the exponential of a block - circulant matrix . in practical computations , we obtain a block - vector with components @xmath566 , @xmath293 , where @xmath567 is bounded in theorem [ th : err2 ] and @xmath568 where , by theorem [ thm : erremb ] , @xmath569 is such that @xmath457\|_\infty\le f_k(\sigma).\ ] ] altogether , for the overall error @xmath570 , one has @xmath571 a similar analysis can be carried out for the relative error . in this case
the inequality @xmath558 is replaced by @xmath572 , for @xmath573 \|_\infty$ ] . so that the function @xmath562 is modified by replacing @xmath64 with @xmath574 . like at the end of section [ sec : epsc ] , in the overall estimate of the error ,
the quantities @xmath472 and @xmath473 are bounded from above by @xmath379 , and we may scale the block - vector @xmath184 so that these quantities are bounded by 1 .
the overall procedure is summarized in algorithm [ alg : ttemb ] , where the repeated squaring of the block - triangular block - toeplitz matrices can be performed by using algorithm [ alg:2 ] , as explained in remark [ rem : tt ] .
set @xmath379 , @xmath485 and @xmath486 set @xmath575 with @xmath576 for @xmath2 , @xmath577 for @xmath383 apply algorithm [ alg : expcirc ] to compute the first block - row @xmath578 of @xmath579 set @xmath580 , for @xmath2 . in this section we use the taylor series method for computing the exponential of an essentially nonnegative matrix , where the block - triangular block - toeplitz structure is exploited to perform fast matrix - vector multiplications .
the computation of the exponential of an essentially nonnegative matrix have been analyzed in @xcite and @xcite .
following @xcite and @xcite , the taylor series method is applied to compute @xmath581 , since the matrix @xmath582 is nonnegative and @xmath62 can be obtained by means of the equation @xmath583 . in this way , we avoid possible cancellations in the taylor summation .
denote by @xmath584 the taylor series truncated at the @xmath585th term , namely @xmath586 the following bound on the approximation error is given in @xcite .
[ thm : trtay ] let @xmath585 be such that @xmath587 .
then @xmath588 the scaling and squaring method is used to accelerate the convergence of the taylor series , by using the property that @xmath589 indeed , if @xmath590 is an estimate of @xmath591 , and if @xmath592 , then @xmath593 and the truncated taylor series expansion is used to approximate @xmath594 . since @xmath595 is block - triangular block - toeplitz , then @xmath596 .
the toeplitz structure is used in the computation of the taylor expansion and in the squaring procedure .
in fact , the computation of each term in the power series expansion consists in performing products between block - triangular block - toeplitz matrices , that can be done by applying algorithm [ alg:2 ] in view of remark [ rem : tt ] ; similarly in the squaring procedure at the end of the algorithm .
concerning rounding errors , we observe that the taylor polynomial is the sum of nonnegative terms .
therefore no cancellation error is encountered in this summation .
the main source of rounding errors is the computation of the powers @xmath597 for @xmath598 which are computed by means of algorithm [ alg:2 ] in view of remark [ rem : tt ] relying on fft .
we omit the error analysis of this computation , which is standard . however , we recall that in view of theorem [ th : fft ] , fft is normwise backward stable but not component - wise stable . for this reason , for the truncation of the power series
it is convenient to replace the component - wise bound expressed by theorem [ thm : trtay ] by the norm - wise bound @xmath599 from which we obtain that the condition @xmath600 implies that @xmath601 .
the overall procedure is stated in algorithm [ alg : ttt ] .
it is worth pointing out that , if the computation of the powers of the triangular toeplitz matrices is performed with the standard algorithm then the computation is component - wise stable as shown in @xcite . set @xmath379 set @xmath602 , @xmath603 , @xmath604 , @xmath130 compute @xmath590 an estimate of @xmath605 , or set @xmath606 compute @xmath592 and @xmath607 set @xmath608 and @xmath237 , @xmath609 , @xmath580 , @xmath130 .
the numerical experiments have been performed in matlab . to compute the error obtained with the proposed algorithms
we have first computed the exponential by using the ` vpa ` arithmetic of the symbolic toolbox with 40 digits and we have considered this approximation as the exact value .
denote by @xmath611 , @xmath612 , the approximations of the blocks on the first row of @xmath62 and define the four errors @xmath613\|_\infty,\\ & \textrm{nw - rel}=\|[a_0-\tilde a_0,\ldots , a_{n-1}-\tilde a_{n-1}]\|_\infty/\|[a_0,\ldots , a_{n-1}]\|_\infty,\\ \end{split}\ ] ] which represent absolute / relative component - wise and norm - wise errors , respectively . the test matrix @xmath59 is taken from two real world problems concerning the erlangian approximation of a markovian fluid queue @xcite .
the block - size @xmath44 of @xmath59 is usually very large since a bigger @xmath44 leads to a better erlangian approximation , while the size @xmath61 of the blocks is equal to 2 for both problems .
we show the performances in terms of accuracy of the algorithm based on the @xmath64-circulant matrix . in table
[ tab:1 ] we report the errors generated by algorithm [ alg : ttepsc ] with @xmath614 applied to the first problem .
observe that the errors are much smaller in magnitude than @xmath73 .
the component - wise and norm - wise absolute errors range around @xmath615 , while the componentwise relative errors deteriorate as @xmath44 increases ; the norm - wise relative errors moderately increase as @xmath44 increases .
this behavior is expected since the use of fft makes the algorithm stable in norm , while the component - wise accuracy is not guaranteed .
concerning the second problem , we report only the results in the case where scaling is applied .
in fact , there is not much differences between the sclaed and the unscaled versions since this problem is already well scaled in its original formulation . in figure
[ fig : prob2.1 ] we report the errors for the method based on @xmath64-circulant matrices .
it is interesting to note that the optimal value of @xmath73 is close to 1 and that the component - wise relative error is minimized by values of @xmath73 greater than 1 .
this fact , which apparently seems to be a contradiction , is explained as follows .
large values of @xmath64 generate large errors in the lower triangular part , i.e. , the lower triangula part of @xmath616 has large norm .
on the other hand we consider the first block - row of @xmath71 to approximate the matrix exponential of @xmath59 , therefore the errors are not influenced by a large error in the lower triangular part . to conclude , the method based on @xmath64-circulant is the fastest one , but the accuracy of the results is lower than that provided by the embedding and taylor series expansion .
however , by applying the averaging technique we can dramatically improve the accuracy of the @xmath64-circulant algorithm .
the algorithms based on embedding , on @xmath64-circulant matrices are faster than the one based on taylor series with fft matrix arithmetic .
moreover they are better suited for a parallel implementation .
a. h. al - mohy , n. j. higham , http://dx.doi.org/10.1007/s11075-009-9323-y[the complex step approximation to the frchet derivative of a matrix function ] , numer .
algorithms 53 ( 1 ) ( 2010 ) 113148 .
http://dx.doi.org/10.1007/s11075-009-9323-y [ ] .
m. benzi , p. boito , http://dx.doi.org/10.1016/j.laa.2013.11.027[decay properties for functions of matrices over @xmath617-algebras ] , linear algebra appl . 456
( 2014 ) 174198 .
[ ] . http://dx.doi.org/10.1016/j.laa.2013.11.027 m. benzi , p. boito , n. razouk , http://dx.doi.org/10.1137/100814019[decay properties of spectral projectors with applications to electronic structure ] , siam rev .
55 ( 1 ) ( 2013 ) 364 .
http://dx.doi.org/10.1137/100814019 [ ] .
d. bini , http://dx.doi.org/10.1137/0213019[parallel solution of certain toeplitz linear systems ] , siam j. comput . 13 ( 2 ) ( 1984 ) 268276 .
d. a. bini , g. latouche , b. meini , http://dx.doi.org/10.1093/acprof:oso/9780198527688.001.0001[numerical methods for structured markov chains ] , numerical mathematics and scientific computation , oxford university press , new york , 2005 , oxford science publications .
n. j. higham , http://dx.doi.org/10.1137/090768539[the scaling and squaring method for the matrix exponential revisited ] , siam rev .
51 ( 4 ) ( 2009 ) 747764 .
http://dx.doi.org/10.1137/090768539 [ ] .
http://dx.doi.org/10.1137/090768539 i. najfeld , t. f. havel , http://dx.doi.org/10.1006/aama.1995.1017[derivatives of the matrix exponential and their computation ] , adv . in appl .
16 ( 3 ) ( 1995 ) 321375 .
http://dx.doi.org/10.1006/aama.1995.1017 m. shao , w. gao , j. xue , http://dx.doi.org/10.1137/120894294[aggressively truncated taylor series method for accurate computation of exponentials of essentially nonnegative matrices ] , siam j. matrix anal .
35 ( 2 ) ( 2014 ) 317338
. http://dx.doi.org/10.1137/120894294 [ ] .
http://dx.doi.org/10.1137/120894294 d. stanford , f. avram , a. badescu , l. breuer , a. da silva soares , g. latouche , phase - type approximations to finite - time ruin probabilities in the sparre andersen and stationary renewal risk models , astin bulletin 35 ( 2005 ) 131144 .
j. xue , q. ye , http://dx.doi.org/10.1090/s0025-5718-2013-02677-4[computing exponentials of essentially non - negative matrices entrywise to high relative accuracy ] , math .
82 ( 283 ) ( 2013 ) 15771596 . http://dx.doi.org/10.1090/s0025-5718-2013-02677-4 [ ] . http://dx.doi.org/10.1090/s0025-5718-2013-02677-4 j. xue , q. ye , http://dx.doi.org/10.1007/s00211-008-0167-5[entrywise relative perturbation bounds for exponentials of essentially non - negative matrices ] , numer .
110 ( 3 ) ( 2008 ) 393403 . | the erlangian approximation of markovian fluid queues leads to the problem of computing the matrix exponential of a subgenerator having a block - triangular , block - toeplitz structure . to this end
, we propose some algorithms which exploit the toeplitz structure and the properties of generators .
such algorithms allow to compute the exponential of very large matrices , which would otherwise be untreatable with standard methods .
we also prove interesting decay properties of the exponential of a generator having a block - triangular , block - toeplitz structure .
* keyword * matrix exponential , toeplitz matrix , circulant matrix , markov generator , fluid queue , erlang approximation . |
let @xmath0 be a knot in the 3-sphere @xmath1 .
a seifert surface @xmath2 for @xmath0 in @xmath1 is said to be _
free _ if the fundamental group @xmath3 is a free group .
we note that all knots bound free seifert surfaces , e.g. canonical seifert surfaces constructed by the seifert s algorithm .
we define the _ free genus _
@xmath4 of @xmath0 as the minimal genus over all free seifert surfaces for @xmath0 ( @xcite ) . for usual genus , schubert ( ( * ? ? ?
* proposition ) ) proved that genus of knots is additive under connected sum . in general , the genus of a knot is not equal to the free genus of it .
in fact , free genus may have arbitrarily high gaps with genus ( @xcite , @xcite ) . in this paper , we show the following theorem . for two knots
@xmath5 , @xmath6 in @xmath1 , @xmath7 .
we can deform a seifert surface @xmath2 by an isotopy so that @xmath8 .
we denote the exterior @xmath9 by @xmath10 , and the exterior @xmath11 or @xmath12 by @xmath13 .
we have the following proposition .
@xmath14@xcite , ( * ? ? ?
* iv.15 ) , ( * ? ? ?
* lemma 2.2.)@xmath15 a seifert surface @xmath2 is free if and only if @xmath13 is a handlebody .
we have the following inequality .
@xmath16 .
let @xmath17 be a free seifert surface of minimal genus for @xmath18 .
we construct a seifert surface @xmath2 for @xmath19 as the boundary connected sum of @xmath20 and @xmath21 naturally
. then @xmath13 is obtained by a boundary connected sum of @xmath22 and @xmath23 .
therefore the exterior of @xmath2 is a handlebody , and @xmath2 is free .
hence we have the desired inequality .
we can specify the _
@xmath24-side _ and _ @xmath25-side _ of a seifert surface @xmath2 for a knot @xmath0 by the orientability of @xmath2 .
we say that a compressing disk @xmath26 for @xmath2 is a _ @xmath24-compressing disk _ ( resp . _ @xmath25-compressing disk _ ) if the collar of its boundary lies on the @xmath24-side ( resp .
@xmath25-side ) of @xmath2 , and @xmath2 is called
_ @xmath24-compressible _ ( resp .
_ @xmath25-compressible _ ) if @xmath2 has a _ @xmath24-compressing disk _ ( resp .
_ @xmath25-compressing disk _ ) .
a seifert surface is said to be _ weakly reducible _ if there exist a @xmath24-compressing disk @xmath27 and a @xmath25-compressing disk @xmath28 for @xmath2 such that @xmath29 .
otherwise @xmath2 is _ strongly irreducible_. the seifert surface @xmath2 is _ reducible _ if @xmath30 .
otherwise @xmath2 is _
irreducible_. a free seifert surface of minimal genus is irreducible .
suppose that @xmath2 is reducible .
then there exist a @xmath24-compressing disk @xmath27 and a @xmath25-compressing disk @xmath28 for @xmath2 such that @xmath30 . by a compression of @xmath2 along @xmath27 ( this is the same to a compression along @xmath28 )
, we have a new seifert surface @xmath31 .
since @xmath32 is homeomorphic to a component of the manifold which is obtained by cutting @xmath13 along @xmath33 , it is a handlebody .
hence @xmath31 is free , but it has a lower genus than @xmath2 .
this contradicts the minimality of @xmath2 . for a free seifert surface @xmath2 of minimal genus for @xmath19 and a decomposing sphere @xmath34 for the connected sum of @xmath5 and @xmath6
, we will show ultimately that @xmath34 can be deformed by an isotopy so that @xmath34 intersects @xmath2 in a single arc , and we have the equality in theorem 1 . to do this
, we divide the proof of theorem 1 into two cases ; ( 1 ) @xmath2 is strongly irreducible , ( 2 ) @xmath2 is weakly reducible .
the case ( 1 ) is treated in the next section and we consider the case ( 2 ) in section 4 .
if a free seifert surface @xmath2 of minimal genus for @xmath19 is incompressible , then an innermost loop argument shows that a decomposing sphere @xmath34 for @xmath19 can be deformed by an isotopy so that @xmath34 intersects @xmath2 in a single arc , and by proposition 3 , we have the equality in theorem 1 .
so , hereafter we suppose that @xmath2 is compressible and that in this section , @xmath2 is strongly irreducible . without loss of generality , we may assume that there is a @xmath24-compressing disk for @xmath2 .
let @xmath35 be a @xmath24-compressing disk system for @xmath2 , and let @xmath31 be a surface obtained by compressing @xmath2 along @xmath35 .
we can retake @xmath35 so that @xmath31 is connected since @xmath13 is a handlebody .
take @xmath35 to be maximal with respect to above conditions .
we deform @xmath31 by an isotopy so that @xmath36 . put @xmath37 , and
let @xmath38 be a closed surface which is obtained by pushing @xmath39 into the interior of @xmath32 .
let @xmath40 be a vertical annulus connecting a core of @xmath41 and a core of the copy of @xmath41 in @xmath38 .
then @xmath38 bounds a handlebody @xmath42 in @xmath32 since @xmath42 is obtained from @xmath13 by cutting along @xmath35 .
the remainder @xmath43 is a compression body since it is obtained from @xmath44 by adding 1-handles @xmath45 .
@xmath31 is incompressible in @xmath1 .
we specify the @xmath46-side of @xmath31 endowed from @xmath2 naturally .
suppose that @xmath31 is @xmath24-compressible , and let @xmath47 be a @xmath24-compressible disk for @xmath31
. then we can regard @xmath47 as a @xmath48-reducing disk for @xmath32 . by applying our situation to (
* lemma 1.1 . )
, we may assume that @xmath49 .
if @xmath50 separates @xmath31 , then @xmath47 cuts off a handlebody from @xmath32 , and there is a non - separating disk in it .
so , we may assume that @xmath50 is non - separating in @xmath31 .
then @xmath51 is a @xmath24-compressing disk system satisfying the previous conditions .
this contradicts the maximality of @xmath35 .
next , suppose that @xmath31 is @xmath25-compressible , and let @xmath52 be a @xmath25-compressing disk for @xmath31 .
then we can regard @xmath52 as a @xmath48-reducing disk for @xmath32 . by applying our situation to ( * ? ? ? * lemma 1.1 . )
, we may assume that @xmath53 is a single loop , and by exchanging @xmath35 if it is necessary , that @xmath52 does not intersect @xmath35 .
but this contradicts the strongly irreducibility of @xmath2 . by lemma 5
, we can deform the decomposing sphere @xmath34 by an isotopy so that @xmath34 intersects @xmath31 in a single arc .
put @xmath54 .
then @xmath55 is a @xmath48-reducing disk for @xmath32 .
otherwise , at least one of @xmath5 or @xmath6 is trivial , and theorem 1 clearly holds . by applying our situation to (
* lemma 1.1 . )
, we may assume that @xmath55 intersects @xmath38 in a single loop , @xmath55 intersects @xmath40 in two vertical arc , and by exchanging @xmath35 under the previous conditions if it is necessary , that @xmath55 does not intersects @xmath35 .
then @xmath34 intersects @xmath2 in a single arc , hence we obtain the inequality @xmath56 .
this and proposition 3 complete the proof of theorem 1 .
in this section , we consider the case that @xmath2 is weakly compressible .
we use the _ hayashi - shimokawa ( hs- ) complexity _ ( @xcite ) . here
we review it .
let @xmath38 be a closed ( possibly disconnected ) 2-manifold .
put @xmath57 is a component of @xmath58 , where this multi - set " may contain the same ordered pairs redundantly .
we order finite multi - sets as follows : arrange the elements of the multi - set in each one in monotonically non - increasing order , then compare the elements lexicographically .
we define the hs - complexity @xmath59 as a multi - set obtained from @xmath60 by deleting all the @xmath61 elements .
we order @xmath59 in the same way as @xmath62 .
since @xmath2 is weakly reducible , there exist @xmath24-compressing disk @xmath27 and @xmath25-compressing disk @xmath28 for @xmath2 such that @xmath29 .
if @xmath65 , say , then @xmath66 bounds a @xmath24-compressing disk for @xmath2 . hence @xmath2 is reducible , and by proposition 4 , a contradiction .
let @xmath76 be a 2-manifold obtained by compressing @xmath2 along @xmath74 , and @xmath31 be a 2-manifold obtained by compressing @xmath2 along @xmath77 .
we deform @xmath78 and @xmath79 by an isotopy so that @xmath80 and @xmath81 . put @xmath37 , and let @xmath38 be a closed 2-manifold which is obtained by pushing @xmath82 into the interior of @xmath32 .
let @xmath40 be a vertical annulus connecting a core of @xmath41 and a core of the copy of @xmath41 in @xmath38 .
then @xmath38 bounds the union of handlebodies @xmath42 in @xmath32 since @xmath42 is obtained from @xmath13 by cutting along @xmath74 .
the remainder @xmath43 is a union of compression bodies since it is obtained from @xmath44 by adding 1-handles @xmath83 .
suppose that there is a 2-sphere component @xmath85 of @xmath38 .
we may assume that @xmath38 does not contain @xmath41 , and there is a copy of some component of @xmath67 in @xmath38 .
let @xmath86 be a subsystem of @xmath67 the union of whose boundaries separates @xmath2 .
if there is no copy of @xmath68 in @xmath85 , then we delete any one of @xmath86 .
then @xmath74 holds the previous conditions , but this contradicts the minimality of @xmath87 .
if there is a copy of @xmath68 in @xmath85 , then there is a simple closed curve in @xmath85 which separates @xmath88 from @xmath89 , and bounds a @xmath24-compressing disk and @xmath25-compressing disk for @xmath2 .
hence @xmath2 is reducible , but this contradicts proposition 4 .
we specify the @xmath46-side of @xmath76 and @xmath31 endowed from @xmath2 naturally .
suppose without loss of generality that @xmath31 is @xmath24-compressible , and let @xmath47 be a @xmath24-compressing disk for @xmath31 .
then we can regard @xmath47 as a @xmath48-reducing disk for @xmath32 . by applying our situation to ( * ? ? ? * lemma 1.1 . )
, we may assume that @xmath47 intersects @xmath38 in a single loop which does not intersect @xmath40 .
we deform @xmath47 by an isotopy so that @xmath90 in @xmath1 .
we take a complete meridian disk system @xmath91 of @xmath92 which includes @xmath67 and does not intersect @xmath47 .
put @xmath93 .
then we have @xmath94 @xmath71 @xmath95 since @xmath50 is essential in @xmath31 .
suppose that @xmath94 @xmath96 @xmath97 .
then each component of @xmath98 bounds both @xmath24-compressing disk and @xmath25-compressing disk for @xmath2 .
hence @xmath2 is reducible , but this contradicts proposition 2.3 .
similarly , if @xmath94 @xmath96 @xmath99 , then we are done .
hence we obtain a @xmath100-compressing disk system @xmath101 , @xmath102 for @xmath2 which satisfies the conditions ( 1 ) , ( 2 ) and have more minimal complexity than @xmath77 .
this contradicts the property of @xmath77 . by lemma 7
, we can deform the decomposing sphere @xmath34 by an isotopy so that @xmath34 intersects @xmath31 in a single arc .
put @xmath54 .
then @xmath55 is a @xmath48-reducing disk for @xmath32 .
otherwise , at least one of @xmath5 and @xmath6 is trivial , and theorem 1 clearly holds .
let @xmath103 and @xmath104 be components of @xmath42 and @xmath92 respectively , where @xmath103 contains @xmath41 and @xmath104 is the next handlebody to @xmath103 .
put @xmath105 .
then @xmath106 gives a heegaard splitting of @xmath107 . by lemma 7
, we can deform @xmath55 by an isotopy so that @xmath55 is contained in @xmath107 . by applying this situation to ( * ? ? ?
* lemma 1.1 ) or ( * ? ? ?
* theorem 1.3 ) , we may assume that @xmath55 intersects @xmath106 in a single loop without moving @xmath108 .
moreover , there exist a complete meridian disk system @xmath109 of @xmath103 such that @xmath110 and @xmath111 .
thus @xmath34 intersects @xmath2 in a single arc , hence we have the conclusion . | we show that free genus of knots is additive under connected sum . |
replacing a non - relativistic wave function in ( [ e5 ] ) by the eigenstates of the dirac equation , one obtains for @xmath94 : @xmath95 we have used the notations of @xcite : @xmath96 and @xmath97 are the radial parts of the large and small components associated with the quantum numbers @xmath98 .
these functions are multiplied by the spherical tensors @xmath99 which are defined by eq .
4a and 4b of @xcite .
the energy of the spin orbit partners of a given value of @xmath37 is denoted by @xmath100 if @xmath29 and @xmath101 if @xmath30 , respectively .
@xmath106 \nonumber \\
\fl \hspace{5ex } + w_{l , m}w_{l-2,m-2 } \left [ f_{-+ } \sqrt{\frac{(l+m)(l-1+m)(l-2+m)(l-3+m)}{(2l-1)^2(2l+1)(2l+3 ) } } \cos{(\omega''_lt ) } \right ] \nonumber \\
\fl \hspace{5ex } + w_{l , m}w_{l , m-2 } \frac{\sqrt{(l+m)(l-1+m)}}{(2l+1 ) } \hspace{-1ex}\\ \times \left [ \frac{\sqrt{(l+1-m)(l+2-m)}}{(2l+1 ) } [ g_{+}+g_{-}-2g_{-+}\cos{(\omega_l t ) } ] \right .
\nonumber \\ \left .
\hspace{3ex } -\frac{\sqrt{(l+2-m)(l-3-m)}}{(2l-1 ) } f_{- } -\frac{\sqrt{(l+2-m)(l+1-m)}}{(2l+3 ) } f_{+ } \right ] \nonumber \\
\fl \hspace{5ex } - w_{l , m}w_{l-2,m } \left[2\ , f_{-+ } \sqrt{\frac{(l^2-m^2)((l-1)^2-m^2)}{(2l-1)^2(2l+1)(2l-3 ) } } \cos{(\omega''_l t ) } \right ] \nonumber \\ \fl \hspace{5ex } \left .
+ w_{l , m}w_{l+2,m-2 } \left [ f'_{-+ } \sqrt{\frac{(l+1-m)(l+2-m)(l+3-m)(l+4-m)}{(2l+1)(2l+3)^2(2l+5 ) } } \cos{(\omega'_l t ) } \right ] \right\ } \nonumber , \end{aligned}\ ] ] @xmath107 \right.\\ \fl \hspace{5ex } - w_{l , m}w_{l-2,m-2 } \left [ f_{-+ } \sqrt{\frac{(l+m)(l-1+m)(l-2+m)(l-3+m)}{(2l-1)^2(2l+1)(2l+3 ) } } \sin{(\omega''_l t ) } \right ] \nonumber \\ \fl \hspace{5ex } + w_{l , m}w_{l-2,m } \left[2\ , f_{-+ } \sqrt{\frac{(l^2-m^2)((l-1)^2-m^2)}{(2l-1)^2(2l+1)(2l-3 ) } } \sin{(\omega''_l t ) } \right ] \nonumber \\ \fl \hspace{5ex } \left . - w_{l , m}w_{l+2,m-2 } \left [ f'_{-+ } \sqrt{\frac{(l+1+m)(l+2-m)(l+3-m)(l+4-m)}{(2l-1)(2l+3)^2(2l+5 ) } } \sin{(\omega'_l t ) } \right ] \right\ } \nonumber , \end{aligned}\ ] ] @xmath108 \nonumber \\
\fl \hspace{5ex } \left .
+ 4(a^2-b^2)\ , w_{l , m}w_{l+2,m}\ , f'_{-+ } \sqrt{\frac{((l+1)^2-m^2)((l+2)^2-m^2)}{(2l+1)(2l+3)^2(2l+5 ) } } \cos{(\omega'_l t ) } \right\ } \end{aligned}\ ] ] in the above formulas , the following notations have been introduced : @xmath109 note that @xmath110 .
radial integrals are denoted as follows : @xmath111 apart from the case of @xmath112 for @xmath113 , which are relatively easily obtained analytically , all other radial integrals have been calculated numerically ( using quadruple precision ) .
the autocorrelation function can be calculated from ( [ a1 ] ) in a straightforward way : @xmath114\right .
\nonumber \\ & & \hspace{2ex } + \left .
\exp{(-ie_l^-t)}\left [ \sum_m w_{l , m}^2 \left ( a^2\frac{l - m}{2l+1}+b^2\frac{l+m}{2l+1 } \right)\right]\right\ } \end{aligned}\ ] ] let us use the following notation : @xmath115 the exact eigenergies @xmath116 ( in units @xmath117 ) are given by @xmath118^{-1/2 } \,.\ ] ] expanding this expression in taylor series with respect to @xmath119 one obtains @xmath120 ^ 8\,.\ ] ] in eq .
[ e8 ] only the lowest term depending on @xmath121 in @xmath122 _ i.e. _ @xmath123 has been included .
the higher order term @xmath124 contributes very little , because the ratio @xmath125 reaches the maximum value about 0.0005 for @xmath15 and @xmath49 and stays much smaller for lower @xmath0 .
then the time evolution for not too long period is mainly determined by the lowest order contribution ( [ e8 ] ) . the precession time is determined by the derivative @xmath126 \;.\ ] ] again in eqs .
( [ e9])-([e10 ] ) only term of the order of @xmath122 has been used .
the @xmath127-order term contributes at most about 0.00022 of the @xmath122-order term for @xmath15 and @xmath49 .
therefore we conclude that the @xmath127-order term can be safely neglected in estimation of the precession time @xmath42 .
99 for an introduction and a list of references see schleich w p 2001 _ quantum optics in phase space _
( berlin , viley - vch . )
bosanac s d 1993 _ j. phys .
a : math . gen . _ * 25 * 5523 klystra n j , ermolayev a m and joachin c j 1997 _ j. phys .
b : at . mol . opt .
phys . _ * 30 * l449 rathe u w , keitel c h , protopapas m and knight p l 1997 _ j. phys .
phys . _ * 30 * l531 rutherford g h and grobe r 1998 _ j. phys . a : math .
gen . _ * 31 * 9931 su q , smetanko b a and grobe r 1998 _ laser phys . _ * 8 * 93 braun j w , su q and grobe r 1999 _ phys .
a _ * 59 * 604 wagner r e , peverly p j , su q and grobe r 1999 _ phys . rev . a _ * 61 * 035402 krekora p , wagner r e , su q and grobe r 2001 _ phys . rev . a _ * 63 * 025404 krekora p , su q and grobe r 2001 _ j. phys .
* 34 * 2795 nilsen h m , masden l b and hansen j p 2001 _ j. phys .
phys . _ * 34 * l39 arvieu r , rozmej p and turek m 2000 _ phys .
rev . a _ * 62 * 022514 arvieu r and rozmej p 1994 _ phys .
a _ * 50 * 4376 ; + 1996 _ j. phys .
phys . _ * 29 * 1339 averbukh i s and perelman n f 1989 _ phys .
lett . a _ * 139 * 449 ; 1989 _ sov . phys .
jetp _ * 69 * 464 rozmej p , arvieu r , averbukh i s and turek
m 2001 ecamp vii , the seventh european conference on atomic and molecular physics , berlin , 2 - 6 april 2001 , europhysics converence abstracts , vol .
25b , p.99 gay j c , delande d and bommier a 1989 _ phys .
a _ * 39 * 6587 crawford m c 2001 _ am .
j. phys . _ * 69 * 1182 chang e c 1985 _ phys
. rev . a _ * 31 * 495 sommerfeld a 1921 _ atombau und spektrallinien _
( f. fievag und sohn , braunschweig ) rozmej p , arvieu r 1999 _ j. phys . a : math .
* 32 * 5637 | time evolution of wave packets built from the eigenstates of the dirac equation for a hydrogenic system is considered .
we investigate the space and spin motion of wave packets which , in the non - relativistic limit , are stationary states with a probability density distributed uniformly along the classical , elliptical orbit ( elliptic wp ) .
we show that the precession of such a wp , due to relativistic corrections to the energy eigenvalues , is strongly correlated with the spin motion .
we show also that the motion is universal for all hydrogenic systems with an arbitrary value of the atomic number @xmath0 .
the detailed study of the time evolution of quantum wave packets ( wps ) in simple atomic or molecular systems has been the object of growing attention for more than ten years @xcite .
most of the previous theoretical studies were done in non - relativistic framework . in the field of relativistic quantum mechanics most of the efforts
have been focused on the problem of the interaction between the atoms and a mixture of static fields with , most of the time , intense laser fields @xcite . under these conditions the use of a relativistic theory is fully justified since the external field is then able to bring considerable energy to the wp . for isolated atoms , however , the use of relativistic dynamics is more questionable , if the wp is followed or observed only during a short period of time . in ref .
@xcite relativistic wave packets , corresponding to circular orbits , have been constructed for hydrogenic atoms with a large @xmath0 , and propagated over time according to the dirac equation .
particular attention was paid to the spin collapse event , _
i.e. _ to the maximum entanglement , in the course of time , of the spin degree of freedom with the spatial ones .
this phenomenon was first shown to take place for a wp in a harmonic oscillator with a spin - orbit force @xcite , where it is periodic .
for this reason it has been called the _ spin - orbit pendulum_. in the dirac equation with a coulomb potential , it is produced by the built - in spin - orbit force and it is not periodic . the time scale where this effect can manifest itself
was discussed in ref .
@xcite , as a function of the charge @xmath0 of the atom and the average principal quantum number @xmath1 of the wp .
this phenomenon is expected to take place on a longer time scale like the other time dependent quantum effects of spreading and of fractional revivals @xcite .
we intend below to complement this work by studying the possible relativistic precession of elliptic wps and by comparing this precession to the spin motion . a preliminary version of the present work was already presented at the ecamp vii conference @xcite .
there are two possible ways to build up a wp on top of a classical elliptic orbit in non - relativistic mechanics .
one of them is by kicking properly a well designed wp , for example a gaussian wp , that is then evolved in time by the free hamiltonian until it spreads above an average classical ellipsis .
this method is not very easy to apply and its main inconvenience is to produce an internal motion within the ellipsis that is able to blur the precession .
therefore , we have preferred a second method , much simpler and even more elegant , which consists of using the coherent wps of ref .
@xcite , which are stationary states of the non - relativistic coulomb problem .
the space probability density of these states was indeed shown to be highly concentrated around a classical bohr - sommerfeld ellipsis . if these states can be adapted to relativistic dynamics their time evolution will doubtlessly be due to relativistic effects , _
i.e. _ due to the fine structure that will act as a perturbation .
let us first show how to adapt these states to a relativistic theory .
we want the spatial part of the large components of the new wp to tend ( in the non - relativistic limit ) toward the state @xmath2 of @xcite which is defined as @xmath3![(l+m)/2 ] ! } \left [ \frac{(l+m)!(l - m)!(2l+1)}{(n - l-1)!(n+l ) ! } \right]^{1/2 } \nonumber \\
\times \left(\sin \frac{\gamma}{2}\right)^{n - m-1 } \left(\cos \frac{\gamma}{2}\right)^{n+m-1}\ , |n , l , m\rangle \;\ ; = \;\ ; \sum_{lm } w_{lm}^{(n)}\,|n , l , m\rangle \ ; . \end{aligned}\ ] ] the probability density @xmath4 is fairly localized onto a bohr orbit with eccentricity @xmath5 and the average angular momentum is @xmath6 where @xmath7 is the principal quantum number of the orbitals @xmath8 which are admixed in ( [ e1 ] ) .
the sum contains @xmath9 values of @xmath10 with @xmath11 even .
the contribution of states with @xmath12 decreases very rapidly with @xmath10 ( more than one order of magnitude for each 2 units of @xmath10 as shown in fig . 1 ) .
the dominant weights are those with @xmath13 and their distribution is nearly gaussian . the relative contribution of states with @xmath12 increases , however , for larger values of the eccentricity parameter @xmath14 .
the larger admixture of these states causes much faster daphasing of the wp .
therefore for illustration of typical precession phenomena he have chosen the case @xmath15 . in order to study the entanglement of the spin degrees of freedom with the orbital ones , it is natural in a non - relativistic theory to start from a product state of an arbitrary spinor @xmath16 with the state @xmath17 .
@xmath18 it may be expanded in eigenstates of total angular momentum @xmath19 with @xmath20 or @xmath21 and @xmath22 or @xmath23 .
@xmath24 in a similar way as in ref .
@xcite , for a circular wp , the state ( [ e5 ] ) is transformed into a four component relativistic state @xmath25 by replacing in ( [ e5 ] ) the non - relativistic states @xmath26 by the eigenstates of the dirac equation with the same quantum numbers . in this manner
the wp gets small components in the most natural way .
the radial parts of the large and of the small components are taken obviously as different ones .
the probability density of the relativistic wp built in this way is represented in fig .
2 .
the time evolution of the wp is produced by introducing the phase factors @xmath27 and @xmath28 as coefficients of states with @xmath29 and @xmath30 with their corresponding eigenvalues .
the four components @xmath31 @xmath32 of the wp at time @xmath33 are given in the [ aa ] with @xmath25 defined as @xmath34 it should be stressed that @xmath25 for @xmath35 is not any more a product state of the form of eq .
( [ e3 ] ) due to the built - in entanglement contained in the solutions of the dirac equation
. however , since the small components of @xmath25 are indeed very small , this defect has no important effect on the magnitude of the initial spin . since the spin - orbit coupling effect , _
i.e. _ spin - orbit pendulum @xcite , manifests itself more efficiently if @xmath36 and @xmath37 are perpendicular to each other , we choose for our discussion : @xmath38 precession of quantum elliptical states in the coulomb field has already been considered in ref .
@xcite , starting also from the same coherent state as done here .
however , the precession was studied only in non - relativistic quantum dynamics as a perturbation effect and no treatment of the spin was attempted .
let us finally discuss the time units relevant to our problem . to the lowest order in terms of the fine structure constant ,
the energy of an eigenstate @xmath39 of the dirac equation can be written ( in a.u . ) for a hydrogenic atom of charge @xmath0 as @xmath40 with @xmath22 or @xmath23 .
the constant energy @xmath41 produces no effect on wp , since it depends only on @xmath7 .
let us define an average time unit @xmath42 ( @xmath43 for precession ) : @xmath44 brackets @xmath45 in ( [ e9])-([e10 ] ) denote average values , @xmath46 is given by ( [ e2 ] ) .
for @xmath47 , @xmath15 precession time @xmath42 ranges from @xmath48s for @xmath49 to @xmath50s for @xmath51 .
@xmath52 in ( [ e10a ] ) denotes the kepler period .
it is necessary to compare @xmath42 to the radiative lifetime @xmath53 of hydrogenic levels .
we will use the estimation of @xmath53 for a single @xmath54 state given in @xcite . in atomic units
it is @xmath55 which was found to agree within 10% with experimental data .
one obtains the universal ratio : @xmath56 which guarantees that the precession of the wave packet takes place long enough time before even a single photon is emitted .
the occurrence of @xmath57 in this ratio is understandable , since @xmath42 is a classical - like quantity , while @xmath53 is proportional to @xmath58 , because it can be expressed as a ratio of a typical energy of the emitted photon to the classical radiation rate corresponding to the orbital motion .
when @xmath59 the linear terms in the expansion of @xmath60 in the power of @xmath37 contribute on the average to @xmath61 in the exponential factors and the wp is expected to restore .
however the terms of the higher order dephase differently the various partial waves , and this leads to a spreading of the wp near the initial shape @xcite .
see the discussion on the magnitude of these terms in [ ab ] .
expression ( 10 ) or ( 11 ) ( with the kepler period @xmath62 ) is also recognized as the classical precession time in the relativistic coulomb problem @xcite .
let us note that @xmath63 is also ( to the first order ) the energy difference between two spin - orbit partners .
therefore the precession time can also be interpreted as the recurrence time of the spin .
hence we should observe a strong correlation between the spatial motion of the density : the precession , and the spin motion .
let us discuss now the dynamics dependence on the atomic number , @xmath0 .
formula ( [ e10 ] ) suggests that the relativistic effects under consideration in this article , depend crucially on @xmath0 . for
sure the highest possible @xmath0 are indeed required to lower @xmath42 as much as possible .
it is interesting to stress nevertheless , that within a very good approximation , a scaling of @xmath0 is possible which leads to the universal behaviour of the wave packet ( [ e1 ] ) .
first of all since @xmath64 , approximated by eq .
( [ e8 ] ) , scales as @xmath65 , _
i.e. _ as @xmath66 , the variable @xmath0 disappears from @xmath67 if we use the reduced time @xmath68 .
the autocorrelation function ( [ ac ] ) which is expressed only in terms of @xmath67 is the simplest quantity which has a universal behaviour , provided the same @xmath7 and @xmath14 are used for all values of @xmath0 .
the other quantities , like the probability density and the the spin expectation values depend on the radial wave functions and radial integrals . in a non - relativistic theory scaling of the radial wave function is elementary , it is obtained by dividing the radial wave fuction by @xmath69 and multiplying the radial variable by @xmath0 .
this is an exact property .
the ratio small / large components of the relativistic solutions scale as @xmath70 , _
i.e. _ roughly like @xmath71 and the radial wave functions depend also in a more complicated manner on @xmath0 . nevertheless , on the whole as seen in fig . 5 below
, the small components contribute a very small part of the probability density even for @xmath49 ( see also fig . 1 of ref
. @xcite ) .
the scaling of the probability density is therefore entirely governed by the large components _ i.e. _ by the non - relativistic theory . in a similar way we have checked that the radial integrals which contribute to the spin expectation values have the same properties : the integrals @xmath72 , and @xmath73 are equal to 1 , within less than 10@xmath74 and the other @xmath75 contribute in a very small manner , also of the order of 10@xmath74 . therefore , the universal behaviour of the wave packet is justified and except for fig . 2 no value of @xmath0
is given .
for longer times , the energy factors omitted in eq .
( [ e8 ] ) which involve higher powers of @xmath0 , play a role and produce a genuine @xmath0 dependence .
those effects will not be discussed here . the probability density of the wave packet with @xmath76 and @xmath77 , with @xmath49 ,
is shown for a set of times up to @xmath59 in fig . 3 and fig . 4 .
the part of the density coming from the small components , shown in fig . 5 , is also concentrated on top of an ellipsis but its magnitude is a thousands times smaller than the total density .
for @xmath78 the density precesses as described classically with a small dispersion .
however the spreading takes place very rapidly for larger @xmath33 and is quite extended for @xmath59 .
the precession of the ellipsis , the recurrence and spreading can be seen in a more condensed manner in the autocorrelation function represented in fig . 6 for three different spin preparations ( spin up , spin down and @xmath77 ) . for small @xmath68
the wp becomes almost orthogonal to its initial parent , for @xmath59 a recurrence occurs but the overlap is only 0.7 .
another peak occurs at @xmath79 but higher frequencies become important and spread the recurrence . for @xmath80
these higher frequencies play a dominant role .
an example of an approximate revival for @xmath81 is displayed in fig .
7 .
fractional revivals @xcite can also be seen to some extent .
two examples of 1/3 and 1/2 revivals are presented in fig .
8 .
the rough independence of the autocorrelation function on the spin preparation requires some explanation .
we can approximate this function by @xmath82 where we have neglected the terms with very small weights @xmath83 .
the contribution of the two sums above is almost the same because @xmath84 has a smooth variation with @xmath37 on the one hand , and because of the exact equality @xmath85 on the other hand .
finally the time evolution of the spin averages is presented in fig .
9 .
this figure exhibits what we can call the relativistic spin - orbit pendulum .
although the wave packet is not prepared initially in a pure state of spin , the impurity is very small ( for @xmath77 one has @xmath86 ) .
as time goes on the spin stays very nearly in the @xmath87 plane and rotates around @xmath88 with period @xmath42 .
its magnitude slowly decreases , however , and for @xmath89 the spin is amost totally entangled with the orbital motion , since the average of its three projections are almost zero . during a period of time that last for about @xmath90 the angular momentum of the spin
is transferred to the orbital motion and therefore the mean trajectory is not planar anymore @xcite .
since the orbital angular momentum of @xmath91 is much higher than @xmath92 this geometrical effect can hardly be represented graphically . from @xmath93
a revival of spin occurs .
the spin rotates and increases its magnitude at the same time .
this event stays even longer than initially .
however the recurrence is only partial .
in conclusion we can say , that for not too long time , the precession and spin motion of the wps are fairly well described by the following approximation : non - relativistic wave functions of the form ( [ e3 ] ) and relativistic energy eigenvalues . from this point of view
the full , computationally very demanding , relativistic approach is unnecessary .
however , this conclusion may be formulated not a priori but only a posteriori .
we would like to stress the richness of the dynamics just described .
indeed , in addition to the relativistic precession of the ellipsis we have obtained for longer times its fractional revivals . during the evolution the spin of the electron is entangled with its orbital motion to various degrees .
all this agree completely with previous results @xcite and @xcite .
however , it was obtained here in a full relativistic calculation . since we have been able to scale the atomic number @xmath0 we have given a universal behaviour to our wp .
it is , however , clear that this scaling is destroyed in real atoms in a more realistic theory which would take into account quantum defects .
their inclusion would also distort the dynamics , for example it would change the precession time , in a way that is out of reach of our simple theory .
as far as purely relativistic effects are concerned , like the importance of the small components or the zitterbewegung , we have found them negligible in the coulomb problem , in contrast to the dirac oscillator @xcite in which they play a major role .
the authors would like to express their gratitude to the two referees who attracted their attention to the interplay between relativistic precession and radiative processes ( eq .
( 12 ) ) in the coulomb problem .
i.s . averbukh thanks also c.r .
stroud , jr .
for useful discussions . |
as the temperature is reduced through @xmath0 k , liquid helium undergoes a phase transition from helium i ( a classical navier stokes fluid ) to helium ii ( a superfluid ) .
the superfluid state persists through to absolute zero at vapour pressure .
the motion of helium ii is well described by the hydrodynamic two - fluid theory of landau & tisza ( 1987 ) . according to this theory ,
helium ii comprises two perfectly mixed fluids , the viscous normal fluid and the inviscid superfluid , of densities @xmath1 and @xmath2 respectively .
the total density of helium ii , @xmath3 , does not vary much with the temperature @xmath4 , whereas @xmath1 and @xmath2 depend strongly on @xmath4 . at absolute zero , @xmath5 k
, the normal fluid component vanishes and helium ii is entirely superfluid @xmath6 , whilst at the lambda point , @xmath7 , the superfluid component is zero @xmath8 and helium ii becomes helium i , which is a classical navier stokes fluid . if helium ii is rotated with angular velocity @xmath9 greater than some small critical value , then vortex filaments appear in the superfluid ( donnelly , 1991 ) .
the circulation around each vortex filament is quantised , in that @xmath10 where @xmath11 is the superfluid velocity field , @xmath12@xmath13sec@xmath14 is the quantum of circulation ( the ratio of plank s constant and the mass of one helium atom ) and @xmath15 is an arbitrary integration path around the axis of the filament . when helium ii rotates the vortices align themselves to the direction of rotation and form a regular configuration with areal density ( number of vortex filaments crossing the unit area perpendicular to the direction of rotation ) @xmath16 ( feynman s rule ) .
the hydrodynamics of the superfluid state is an interesting topic _ per se _ , but it is worth mentioning that attention to this problem has additional motivations .
the first arises from the engineering applications : helium is the only substance available in liquid form at temperatures near absolute zero , so it is important as a cryogenics coolant .
applications range from infrared detectors in space science to the cooling of superconducting magnets in particle physics .
the second motivation comes from recent experimental developments in which the relation between classical and quantum turbulence is investigated ( for example see smith 1993 ; barenghi , swanson & donnelly 1995 ) .
the most generally accepted equations for modelling the macroscopic flow of helium ii are the hall - vinen - bekharevich - khalatnikov ( hvbk ) equations which were derived by a number of people over the years ( hall & vinen 1956 ; hall 1960 ; bekharevich & khalatnikov 1965 ; hills & robert 1977 ) .
these equations extend landau s two - fluid model to take into account the presence of quantised vortex lines in the flow .
the derivation of the equations is based on a continuum approximation , assuming a high density of vortex lines , all aligned roughly in the same direction . for such situations ,
the superfluid vorticity , which is discrete in nature , may be approximated as a continuum , resulting in an effective superfluid vorticity field @xmath17 .
the isothermal , incompressible hvbk equations may be written as : @xmath18 @xmath19 @xmath20 where @xmath21 is the normal fluid velocity , @xmath22 is the kinematic viscosity of the normal fluid and @xmath23 , @xmath24 are effective pressures ( @xmath25 and @xmath26 where @xmath27 is the pressure ) .
the mutual friction force may be written as @xmath28 with @xmath29 the unit vector in the direction of superfluid vorticity and @xmath30 , @xmath31 are the temperature - dependent mutual friction parameters ( barenghi , donnelly & vinen 1983 , donnelly & barenghi 1998 ) .
this force is due to collisions between the normal fluid ( mainly rotons , at temperatures relevant to most experiments ) and vortex lines . the vortex tension force may be written as @xmath32 and reflects the energy per unit length in the vortex lines .
the vortex tension parameter @xmath33 has the same dimension as kinematic viscosity but physically it is very different : it represents the ability of a vortex line to oscillate due to vortex waves which can be excited on the vortex lines themselves .
the quantity @xmath34 represents the intervortex spacing and @xmath35 cm is the radius of the superfluid vortex core .
in this paper we apply the hvbk equations to model helium ii in a finite aspect ratio couette annulus , that is flow of fluid confined radially between two concentric rotating cylinders and axially between two fixed plates separated by a distance @xmath36 .
we consider the inner cylinder ( @xmath37 ) to be rotating with angular velocity @xmath38 and the outer cylinder ( @xmath39 ) to be fixed , see figure [ fig : couette ] .
couette flow has been used as a bench - mark for fluid mechanics since taylor s pioneering work ( 1923 ) to investigate the transition from couette flow to taylor vortices , which established a firm ground for using the navier
stokes equations and the no - slip boundary conditions .
progress in helium ii has been slower than for classical fluids due in part to problems of flow visualisation at such low temperatures . in considering a classical fluid , introduction of flakes or other small particles into the working fluid ( usually oil or water ) , results in the taylor vortices being clearly evident .
in contrast there are only limited visualisation techniques available to the experimentalist at temperatures close to absolute zero .
recent attempts have been made to reveal the flow pattern of helium ii by adding small particles ( bielert & stamm 1993 ) .
however this was only successful at high rotation rates ( 40 times the critical angular velocity at which linear stability analysis predicts couette flow becomes unstable ) ; in this regime the flow is turbulent and the validity of the hvbk equations is not clear . clearly the lack of direct flow visualisation gives additional motivation to our work .
experiments on helium ii between concentric cylinders were first performed by kapitza in 1941 and donnelly & lamar ( 1988 ) have written a review of experiments involving helium ii in couette apparatus .
early taylor
couette experiments were concerned with determining the viscosity of helium ii by measuring the torque exerted by the flow on the stationary cylinder . a break in the linear dependence of the torque with the angular velocity of the rotating cylinder
is taken to denote a transition from one solution to another .
a second experimental technique is available , that of measuring the extra attenuation of a second sound wave , which can be used to probe the superfluid vorticity .
second sound waves occur when there is a periodic counterflow between the normal fluid and superfluid , which corresponds to a wave of heat .
angular velocity is plotted against the attenuation factor and breaks in the curve are interpreted as transitions in the flow . by measuring the extra attenuation of second sound waves in the axial , azimuthal and radial directions it is theoretically possible to get an idea of the number and direction of the quantised vortex lines . in practice
the information obtained is less complete than this . despite the fact that the hvbk equations were established in 1961
, they have been used mainly to model helium ii in a rotating container ( solid body rotation ) .
unfortunately solid body rotation ( which corresponds to a spatially uniform configuration of superfluid vortices ) is not a strict test of the equations of motion because it is too simple ; the terms in the hvbk equations which involve the vortex tension vanish .
only recently have the hvbk equations been validated for a non - trivial flow configuration : couette flow . for the case of infinitely long cylinders , both superfluid and normal fluid velocity fields have the following couette flow profile : @xmath40 where @xmath41 are cylindrical coordinates and @xmath42 is the unit vector in the azimuthal direction . after making the usual assumption of infinitely long cylinders , barenghi & jones ( 1988 ) and barenghi ( 1992 ) considered the linear stability of couette flow with respect to infinitesimal perturbations of the form @xmath43 where @xmath44 is the complex growth rate , @xmath45 the azimuthal wavenumber and @xmath46 the dimensionless axial wavenumber ( expressed in units of @xmath47 where @xmath48 is the gap width ) .
the driving parameter of the problem is the reynolds number , @xmath49 which represents the dimensionless velocity of the inner cylinder .
barenghi & jones ( 1988 ) found that if the inner cylinder rotates sufficiently fast , at a certain critical value @xmath50 the growth rate , real@xmath51 of the axisymmetric ( @xmath52 ) perturbation becomes positive , hence couette flow becomes unstable .
this transition corresponds to the onset of taylor vortices for a classical fluid .
barenghi & jones ( 1988 ) also determined the temperature dependence of the critical reynolds number , @xmath53 , and of the critical axial wavenumber , @xmath54 .
their work prompted further experiments and good agreement between the predicted and measured values of @xmath53 was found ( barenghi 1992 ) , particularly for temperatures close to the lambda point .
the success of the linear stability analysis prompted further work , namely the study of nonlinear taylor flow in infinitely long cylinders ( henderson , barenghi & jones 1995 ) .
comparisons with existing experimental data and numerical results further validated the hvbk model in the high temperature regime ( henderson & barenghi 1994 ) .
unfortunately at lower temperatures ( @xmath55 k ) , although there was qualitative agreement , the measured critical reynolds numbers were larger than those predicted by the linear stability analysis . in the classical taylor
couette problem the critical value of the dimensionless axial wavenumber is @xmath56 , hence each individual taylor vortex cell is approximately square ( the extension in the axial direction is equal to the gap s size ) . in the case of helium ii , barenghi & jones ( 1998 ) found that @xmath57 as @xmath58 , as expected in the limit of a pure normal fluid .
however , as the temperature is reduced , they found that @xmath59 decreases and tends to zero in the limit of a pure superfluid .
this result suggests that the discrepancy between theory and experiments for @xmath55 k is due to end effects : there are not enough taylor vortex cells in the typical experimental apparatus , thus the infinite cylinder assumption breaks down .
a better theory is required and that is what we set out to do in this paper .
it is known in the classical taylor couette literature that the presence of fixed ends induces a large scale ekman circulation in which the fluid moves radially inward near the top and bottom ends of the couette apparatus and moves radially outward in the middle .
this effect is caused by the no - slip boundary conditions ; the centrifugal force pushes the fluid outwards at the centreline , where the braking effect of the end plates is least , so the fluid near the end plates moves inwards to conserve mass . in the case of helium ii the normal
fluid obeys the same no - slip boundary condition of an ordinary classical fluid , but there are also superfluid boundary conditions to take into account .
a discussion of what these superfluid boundary conditions should be is contained in our previous paper ( henderson & barenghi 2000 ) , in which we also showed that the competition between the normal fluid and superfluid boundary conditions has unexpected effects on the direction of rotation of the ekman circulation in a unit aspect ratio annulus .
the aim of the current paper is to understand how the appearance of taylor vortex flow is effected by the underlying ekman circulation .
firstly we shall investigate the low reynolds number flow of helium ii at varying aspect ratios to see how it differs from the flow of a classical navier stokes fluid .
secondly , we shall determine the transition from ekman cells to taylor vortices and compare the taylor
couette flow of helium ii in an enclosed annulus to the flow in an infinite apparatus ( henderson 1995 ) .
if , for sufficiently long cylinders , the results of the linear stability theory are recovered , contact will be made between theory and experiments at low temperatures .
we consider the fluid to be confined radially between two concentric cylinders of inner and outer radius @xmath60 and @xmath61 , and axially between two fixed plates which are separated by a distance @xmath36 .
the top and bottom plates and the outer cylinder are held stationary and the inner cylinder rotates at constant angular velocity @xmath38 . throughout this work we shall consider the radius ratio , @xmath62 as in the experimental apparatus of swanson & donnelly ( 1991 ) and will vary the reynolds number , @xmath63 , the aspect ratio , @xmath64 and the temperature , @xmath4 . in order to solve the hvbk equations ( [ eq : vneqn]-[eq : div ] ) we need boundary conditions .
the boundary conditions for the normal fluid are no - slip , that is , working in cylindrical coordinates @xmath65 at @xmath37 and @xmath66 at @xmath67 @xmath68 .
the boundary conditions for the superfluid are more delicate . given @xmath69 , a normal to the boundary , we have that @xmath70 at @xmath71 and @xmath72 , ensuring no flow normal to the boundaries . for the remaining conditions we have taken @xmath73 where @xmath74 is the unit vector in the axial direction .
thus the superfluid vorticity is taken to be purely axial at the boundaries .
the condition on the cylinder walls , @xmath37 and @xmath39 , has been discussed in a previous paper ( henderson 1995 ) whilst the condition on the ends of the cylinder , @xmath75 and @xmath76 , corresponds to perfect sliding of the vortex lines ( khalatnikov 1965 ) .
it must be stressed that as yet there is no experimental evidence for the boundary conditions which we propose ; in fact , a motivation behind or work is to explore the consequence of assuming certain boundary conditions , hoping to stimulate experimental work on this issue .
other possibilities are that the vortex lines remain totally or partially pinned at the boundaries .
total pinning of the vortex lines on the ends of the cylinders can not occur because each rotation would wrap up the vortex lines until de - pinning takes place .
partial sliding is a possibility , however this would introduce extra unknown parameters into the problem depending on the nature and the smoothness of the boundaries , so we have chosen to adopt perfect sliding . we make the simplifying assumption that the flow is axisymmetric and introduce stream functions @xmath77 , @xmath78 such that v^p_r&&=- , v^p_z= [ eq : psi ] where @xmath79 , @xmath80 represents the normal fluid , superfluid respectively .
introducing the stream functions ensures that continuity ( [ eq : div ] ) is automatically satisfied .
a finite difference approach is used to obtain solutions for @xmath81 , @xmath82 and @xmath83 .
the equations satisfied by these quantities are obtained by taking the @xmath84 component of ( [ eq : vneqn],[eq : vseqn ] ) and the @xmath84 component of the curl of ( [ eq : vneqn],[eq : vseqn ] ) .
a poisson s equation links the stream function @xmath81 with the azimuthal component of vorticity @xmath83 .
these equations are solved on the computational domain @xmath85 , @xmath86 and symmetry is assumed to calculate the derivatives on the centreline @xmath87 .
a uniform grid is used in both the radial and axial direction and the equations are stepped forward in time until a steady solution is achieved . the computational cost given our criterion of convergence determines the largest aspect ratio which we compute ( @xmath88 ) .
the equations are made dimensionless using the gap width @xmath89 as the unit of length and the normal fluid viscous time scale @xmath90 as the unit of time .
for further details of the numerical method , see henderson & barenghi ( 2000 ) .
in this section we refer to previous results ( henderson & barenghi 2000 ) in which we considered the low reynolds number flow of helium ii in a unit aspect ratio couette annulus ; this geometry constrains the flow to a simple ekman circulation .
the key discovery was the anomalous ekman motion of helium ii , compared to a classical navier stokes fluid .
we found that the pair of superfluid ekman cells always rotate in a counter - classical direction ( which means for example that @xmath91 at the centreline @xmath87 ) , whilst the normal fluid ekman cells rotate classically ( @xmath92 at @xmath87 ) at temperatures close to the lambda temperature , but reverse ( @xmath93 at @xmath87 ) at lower temperatures .
it was also found that the azimuthal superfluid velocity is almost independent of @xmath94 .
this effect becomes more pronounced at lower temperatures and was found to be due to the tension in the superfluid vortex lines .
in this paper , the parameters which are varied are the temperature @xmath4 , reynolds number @xmath95 of the inner cylinder and aspect ratio @xmath96
. figures [ fig : fig2][fig : fig4],[fig : fig6 ] show contour plots of the normal and superfluid stream functions , where the plots are shown for the complete cross - section of the annulus ; @xmath85 , @xmath97 with the inner / outer cylinder on the left / right respectively .
the maximum value of each field is printed underneath the corresponding contour plot .
light and dark regions correspond to positive and negative contour lines respectively . in figure
[ fig : fig2 ] we investigate how varying the aspect ratio effects the flow of helium ii at @xmath98 k and @xmath99 .
this reynolds number is appreciably below the critical reynolds number @xmath100 at which the linear stability analysis predicts couette flow becomes unstable in the infinite cylinder approximation .
it is apparent from figure [ fig : fig2](_a _ ) that since the normal fluid and superfluid stream functions increase with @xmath94 across the centreline @xmath87 , ( [ eq : psi]_a _ ) predicts that the radial velocity components are negative at the centreline , in contrast to the motion of a classical navier stokes fluid .
thus for small aspect ratios ( @xmath101 ) both the normal fluid and superfluid rotate counter - classically in a pair of matching ekman cells .
this behaviour has been reported previously ( henderson & barenghi 2000 ) and is due to the boundary conditions satisfied by the vortex lines on the top and bottom of the cylinders .
figure [ fig : fig2 ] also shows that the superfluid ekman cells fill the whole annulus , becoming more elongated as the aspect ratio is increased .
in contrast , the normal fluid ekman cells , tend to retain their size and are positioned towards the ends of the cylinders .
as the aspect ratio is increased further ( @xmath102 ) the superfluid remains in the form of two counter - rotating cells whilst the normal fluid splits into four cells with the cells adjacent to the ends of the cylinder rotating in the same direction as the superfluid .
the inner cells of the normal fluid strengthen and fill the annulus as the aspect ratio is increased . at aspect ratio @xmath88
the superfluid splits into four cells and although the size of the cells is different to those of the normal fluid , the direction of rotation is the same for both fluids . taken together these results demonstrate that the low - reynolds number flow of helium ii is dominated by end effects , which force the outer cells of the normal and superfluid to rotate counter - classically . as the aspect ratio
is increased the influence of the ends at the centre of the apparatus becomes weaker and classical behaviour is observed in both fluids , with outflow at the centreline , @xmath87 . in our previous work ( henderson & barenghi 2000 ) we highlighted the columnar motion of the azimuthal component of superfluid velocity .
we find that as the aspect ratio is increased this column like motion becomes less pronounced . in figure
[ fig : fig3 ] we show how varying the temperature effects the flow of helium ii at @xmath99 and @xmath103 . at this aspect ratio ,
the maximum axial wavelength of a disturbance is , @xmath104 giving a corresponding minimum admissible axial wavenumber of @xmath105 . given this , the reynolds number , @xmath99 is below the critical value predicted by linear stability analysis at which couette flow become unstable for all temperatures considered . at temperatures close to the lambda temperature ( @xmath106k )
the normal fluid and superfluid rotate in a pair of ekman cells .
the normal fluid rotates in a classical direction and the superfluid rotates counter - classically . as the temperature
is reduced the normal fluid splits into four cells , whilst the superfluid remains as two cells . in this region ,
the normal fluid cells closest to the ends of the cylinder rotate in the same direction as the superfluid and become stronger in magnitude and also fill more of the annulus as the temperature is reduced .
the columnar behaviour of the azimuthal superfluid velocity is not evident at temperatures close to the lambda temperature but becomes more pronounced as the temperature is reduced . in figure [ fig : fig4 ]
we investigate the transition from ekman flow to taylor vortex flow at @xmath107 k and @xmath108 . at low reynolds numbers ( @xmath99 )
the normal fluid consists of a pair of classically rotating ekman cells .
the superfluid has four cells , with the cells adjacent to the ends of the cylinders rotating counter - classically , as predicted by the boundary conditions .
as the reynolds number is increased both the normal fluid and the superfluid develop more cells , with the cells closest to the ends rotating classically and counter - classically for the normal and superfluid respectively , as a consequence of the boundary conditions . at @xmath109 taylor vortex flow
is fully developed and we can see that despite the superfluid having an additional weak pair of cells close to the ends of the cylinders , the two fluids appear to match each other . in figure [
fig : fig5 ] we plot the centreline average values of the radial components of the normal fluid ( solid ) and superfluid ( dashed ) against reynolds number . from the graph
we can see that the transition to taylor vortices occurs at approximately @xmath110 , whilst stability analysis in the infinitely long cylinders approximation predicts @xmath111 at @xmath112 . in figure
[ fig : fig6 ] we repeat the calculation at @xmath108 but setting a lower temperature , @xmath113 k. at low reynolds numbers ( @xmath99 ) the normal fluid has a pair of classically rotating ekman cells .
as before , the superfluid has four cells , and the cells which are adjacent to the ends of the cylinders rotate counter - classically , as predicted by the boundary conditions .
as the reynolds number is increased , the normal fluid develops four cells which fill the whole apparatus , not six as at @xmath107 k. the outer cells rotate classically due to the boundary conditions . the superfluid remains as four cells up to @xmath109 , with the outer cells rotating counter - classically and the inner cells strengthing with increasing reynolds number . at @xmath114
the superfluid develops an additional small weak pair of cells at the centre of the apparatus . in figure
[ fig : fig7 ] we plot the centreline average values of the radial components of the normal fluid ( solid ) and superfluid ( dashed ) against reynolds number .
the transition from the ekman flow to the taylor vortex flow is harder to distinguish than at higher temperatures and is slightly different for the normal and superfluid .
it occurs at approximately @xmath115 .
what is interesting about these results is that the taylor vortices for the normal fluid are elongated , being @xmath116 times longer than they are wide .
this is consistent with the linear stability analysis which predicts that @xmath117 at the critical axial wavenumber of @xmath118 .
the elongation of the taylor cells was first predicted by barenghi & jones ( 1988 ) and this is the first time that it has been observed numerically taking end effects into account .
we find that in general the superfluid vorticity is primarily axial and concentrated towards the ends of the cylinders close to the inner cylinder .
this predicted result could be observed experimentally by measuring the extra attenuation of second sound waves due to the vortex lines at the centre and ends of the apparatus .
as the reynolds number is increased , there is a much larger deflection of the vortex lines in the radial direction .
this deflection has been observed experimentally in the transition to taylor vortices ( swanson & donnelly 1991 ) . in figure
[ fig : fig8 ] we plot the superfluid vorticity components of helium ii at @xmath113 k , @xmath108 at the reynolds numbers ( a ) @xmath99 and ( b ) @xmath109 for the middle two - thirds of the apparatus ( @xmath119 ) .
this enables us to see what happens to the vortex lines in the centre of the apparatus away from the cylinder ends , where there is stronger shear which would make the contour lines in the middle of the apparatus less visible .
the deflection in the radial direction is clearly visible for @xmath109 . in the taylor vortex flow regime ,
the vorticity pattern is similar to that found when considering the taylor
couette flow of helium ii without end effects ( henderson 1995 ) .
by solving the nonlinear hvbk equations in a finite aspect ratio configuration we have determined how the fixed ends of the cylinders effect the ekman flow and the transition to taylor vortex flow . for low - reynolds number flow ,
the main result is the anomalous ekman circulation of helium ii when compared to that of a classical navier stokes fluid . at short aspect ratios both normal
fluid and superfluid rotate as a pair of ekman cells
. the superfluid cells always rotate counter - classically , whilst the sense of rotation of the normal fluid is temperature dependent . at larger aspect ratios the influence of the ends of the cylinders towards the centre of the apparatus diminishes and at lower temperatures the normal fluid develops additional cells resulting in classical outflow at the centreline .
the superfluid also develops extra cells to match the flow of the normal fluid at the centre .
we have found that at @xmath107 k and aspect ratio as small as @xmath108 , in the regions way from the ends of the cylinders , the superfluid s pattern matches that of the normal fluid .
the transition to taylor vortex flow occurs at a reynolds number close to that predicted by the linear stability analysis in the infinite cylinder approximation . at the slightly lower temperature of @xmath113
k we find that the normal fluid develops elongated taylor cells , as predicted by the linear stability analysis .
the superfluid flow pattern is similar to that of the normal fluid away from the ends of the cylinders but the matching is not as pronounced as at higher temperatures . | by numerically computing the steady axisymmetric flow of helium ii confined inside a finite aspect ratio couette annulus , we determine the transition from ekman flow to taylor vortex flow as a function of temperature and aspect ratio .
we find that the low - reynolds number flow is quite different to that of a classical fluid , particularly at lower temperatures . at high aspect ratio our results confirm the existing linear stability theory of the onset of taylor vortices , which assumes infinitely long cylinders . |
to see the additivity of @xmath74 , we reformulate it as @xmath153 , where @xmath154 where @xmath47 is the projection onto @xmath48 .
the dual sdp of @xmath155 is given by @xmath156 the optimal values of the primal and the dual sdps above coincide by strong duality .
for any two bipartite states @xmath157 and @xmath158 , we have that @xmath159 on one hand , suppose that the optimal solution to sdp ( [ dual eta ] ) of @xmath160 and @xmath161 are @xmath162 and @xmath163 , respectively .
it is easy to see that @xmath164\ge 0,\\ y_1\ox y_2-p_1^{t_b}\ox p_2^{t_{b'}}=\frac{1}{2}[(y_1+p_1^{t_b})\ox ( y_2-p_2^{t_{b'}})+(y_1-p_1^{t_b})\ox ( y_2+p_2^{t_{b'}})]\ge 0.\\\end{aligned}\ ] ] then , we have that @xmath165
. moreover , @xmath166 which means that @xmath167 . therefore , @xmath168 is a feasible solution to the sdp ( [ dual eta ] ) of @xmath169 , which means that @xmath170 on the other hand , suppose that the optimal solutions to sdp ( [ prime eta ] ) of @xmath160 and @xmath161 are @xmath171 and @xmath172 , respectively .
assume that @xmath173 it is easy to see that @xmath174 and @xmath175 .
thus , @xmath176 is a feasible solution to the sdp ( [ prime eta ] ) of @xmath169 .
this means that @xmath177 hence , combining eq .
( [ eta le ] ) and eq .
( [ eta ge ] ) , it is clear that @xmath178 , which means that @xmath159
for any @xmath1 with spectral decomposition @xmath114 where @xmath179 it holds that @xmath180 suppose that @xmath181 and @xmath182 . noting that @xmath183 , we have @xmath184 .
thus , with simple calculation , it is easy to see that @xmath185/\sqrt2.\ ] ] then , one can simplify @xmath186 to @xmath187/\sqrt2 \ ( 0\le\theta\le \pi/2),\ ] ] where @xmath188 is determined by @xmath189 and @xmath190 .
we assume that @xmath191 .
moreover , let us choose @xmath196 it is clear that @xmath197 and it can be easily checked that @xmath198 .
thus , @xmath122 is a feasible solution to the sdp ( [ dual eta ] ) of @xmath74 , which means that @xmath199 | we demonstrate the irreversibility of asymptotic entanglement manipulation under quantum operations that completely preserve positivity of partial transpose ( ppt ) , which resolves a major open problem in quantum information theory . to be more specific ,
we show that for any rank - two mixed state supporting on the @xmath0 antisymmetric subspace , the amount of distillable entanglement by ppt operations is strictly smaller than one entanglement bit ( ebit ) while its entanglement cost under ppt operations is exactly one ebit . as a byproduct
, we find that for this class of quantum states , both the rains bound and its regularization , are strictly less than the asymptotic relative entropy of entanglement with respect to ppt states .
so in general there is no unique entanglement measure for the manipulation of entanglement by ppt operations .
we further present a feasible sufficient condition for the irreversibility of entanglement manipulation under ppt operations . *
introduction : * a basic feature of many physical processes is the appearance of irreversibility . in the area of quantum information science , entanglement plays a central role and the irreversibility in entanglement manipulations has been studied in the last two decades .
the manipulation of entanglement under local operations and classical communication ( locc ) is generally irreversible in the finite - copy regime .
more precisely , the amount of pure entanglement that can be distilled from a finite number of copies of a state @xmath1 is usually strictly smaller than the amount of pure entanglement needed to prepare the same number of copies of @xmath1 @xcite .
surprisingly , in the asymptotic settings where the number of copies tend to infinite , this process of entanglement manipulation for bipartite pure states is shown to be reversible @xcite . in contrast , for mixed states , this asymptotic reversibility under locc operations does not hold anymore @xcite .
in particular , one requires a positive rate of pure state to generate the so - called bound entanglement by locc @xcite , while it is well known that no pure state can be distilled from it @xcite .
various approaches have been considered to enlarge the class of operations to ensure reversible interconversion of entanglement in the asymptotic setting .
a natural candidate is the class of quantum operations that completely preserve positivity of partial transpose ( ppt ) @xcite .
a remarkable result is that any state with a nonpositive partial transpose ( npt ) is distillable under this class of operations @xcite .
this suggests the possibility of reversibility under ppt operations and there are examples of mixed states which can be reversibly converted into pure states in the asymptotic setting , e.g. the class of antisymmetric states of arbitrary dimension @xcite .
however , the reversibility under ppt operations remained unknown so far since there were no further examples .
recently , a reversible theory of entanglement considering all asymptotically non - entangling transformations was studied in refs . @xcite and
the unique entanglement measure is identified as the asymptotic ( regularized ) relative entropy of entanglement . a more general reversible framework for quantum resource theories
was introduced in ref .
@xcite .
when the pure entanglement is set to be the standard @xmath2 maximally entangled state , two fundamental ways of entanglement manipulation are well known , namely , entanglement distillation and entanglement dilution @xcite .
these two tasks also naturally raise two fundamental entanglement measures of distillable entanglement @xmath3 and entanglement cost @xmath4 @xcite . to be specific ,
@xmath3 is the highest rate at which one can obtain @xmath2 maximally entangled states from the given state by locc @xcite while @xmath4 is the optimal rate for converting @xmath2 maximally entangled states to the given state by locc .
then the problem of the reversibility in entanglement manipulation under ppt operations is to determine whether distillable entanglement always coincides with entanglement cost under ppt operations .
this problem is very difficult since for general mixed states it is highly nontrivial to evaluate these two measures both of which are given by a limiting procedure .
note that ppt bound entanglement is not a candidate for irreversibility anymore since it can be used as free resources under ppt operations . in this paper
, we demonstrate that irreversibility still exists in the asymptotic entanglement manipulation under ppt operations , which resolves a major open problem in quantum information theory @xcite .
our approach is to show a gap between the regularized rains bound and the asymptotic relative entropy of entanglement with respect to ppt states , which also resolves another open problem in ref .
@xcite . to be specific ,
we introduce an additive semidefinite programming ( sdp ) @xcite lower bound for the asymptotic relative entropy of entanglement with respect to ppt states and compare it to the rains bound . with this
new lower bound , we are able to show that the ppt - assisted entanglement cost of any rank - two state supporting on the @xmath0 antisymmetric subspace is exactly one entanglement bit ( ebit ) while its ppt distillable entanglement is strictly smaller than one ebit . as an immediate corollary , we show that there is no unique entanglement measure under ppt operations .
finally , we give a sufficient condition which can be used to efficiently verify the irreversibility of asymptotic entanglement manipulation under ppt operations .
then a more general class of sates can be constructed to further illustrate the irreversibility .
before we present our main results , let us review some notations and preliminaries .
in the following we will frequently use symbols such as @xmath5 ( or @xmath6 ) and @xmath7 ( or @xmath8 ) to denote ( finite - dimensional ) hilbert spaces associated with alice and bob , respectively .
the set of linear operators over @xmath5 is denoted by @xmath9 .
note that for a linear operator @xmath10 over a hilbert space , we define @xmath11 , and the trace norm of @xmath10 is given by @xmath12 , where @xmath13 is the complex conjugate of @xmath10 .
the operator norm @xmath14 is defined as the maximum eigenvalue of @xmath15 . a deterministic quantum operation ( quantum channel )
@xmath16 from @xmath6 to @xmath7 is simply a completely positive and trace - preserving ( cptp ) linear map from @xmath17 to @xmath18 .
the choi - jamiokowski matrix of @xmath16 is given by @xmath19 , where @xmath20 and @xmath21 are orthonormal basis on isomorphic hilbert spaces @xmath5 and @xmath6 , respectively .
a positive semidefinite operator @xmath22 is said to be ppt if @xmath23 , where @xmath24 means the partial transpose over the system @xmath7 , i.e. , @xmath25 .
a bipartite operation is said to be a ppt operation if its choi - jamiokowski matrix is ppt .
a well known fact is that the classes of ppt operations , separable operations ( sep ) @xcite and locc obey the following strict inclusions @xcite , @xmath26 let @xmath27 represent one of .
then the concise definitions of distillable entanglement and entanglement cost by the class of @xmath27 operations can be given as follows @xcite : @xmath28 where @xmath29 represents the standard @xmath30 maximally entangled state .
for simplicity , we denote @xmath31 and @xmath32 as @xmath33 and @xmath34 , respectively . in ref .
@xcite , hayden , horodecki and terhal proved that @xmath34 equals to the regularised entanglement of formation @xmath35 @xcite , the convex roof extension of the entropy of entanglement .
the rains bound is introduced in ref .
@xcite to upper bound @xmath36 and is reformulated in ref .
@xcite as the following convex optimization problem : @xmath37 in this formula , @xmath38 denotes the quantum relative entropy , where we take @xmath39 throughout the paper . since rains bound is not additive @xcite , it is necessary to consider the regularized rains bound , i.e. , @xmath40 the relative entropy of entanglement ( ree ) @xcite with respect to the ppt states is given by @xmath41 and the asymptotic relative entropy of entanglement with respect to the ppt states is given by @xmath42 the logarithmic negativity @xcite is an upper bound on ppt - assisted distillable entanglement , i.e. , @xmath43 .
an improved version of logarithmic negativity is introduced in ref . @xcite as @xmath44 it was shown in ref .
@xcite that @xmath45 the following one - copy ppt - assisted deterministic distillable entanglement was also obtained in ref .
@xcite , @xmath46 where @xmath47 is the projection onto @xmath48 , the support of @xmath1 .
note that @xmath48 is defined to be the subspace spanned by the eigenvectors of @xmath1 with positive eigenvalues .
clearly @xmath49 is efficiently computable by sdp , and we have @xmath50 which is very helpful to determine the exact values of ppt - assisted distillable entanglement and the rains bound for some states .
the entanglement distillation and dilution of bipartite states under locc / sep / ppt operations satisfy the following chain of inequalities : @xmath51 where the fifth inequality is proved in ref .
@xcite .
* irreversibility under ppt operations : * we are going to show the irreversibility under ppt operations by using a @xmath52 state , namely , @xmath53 with @xmath54 the projection onto the support of @xmath55 is given by @xmath56 .
it is clear that the state @xmath55 is a rank - two state supporting on the @xmath52 antisymmetric subspace . in ref .
@xcite , chitambar and one of us showed that this state can be transformed into a @xmath2 pure entangled state by a suitable separable operation while no finite - round locc protocol can do that . for the state @xmath55 , @xmath57 from the propositions [ ppt cost ] and [ ppt distill ] below ,
the gap between the regularized rains bound and the asymptotic relative entropy of entanglement is clear to see . in particular ,
the irreversibility under ppt operations is directly proved by @xmath58 * sdp lower bound for @xmath59 : * our aim is to obtain an efficiently computable lower bound for @xmath59 . in the one - copy case
, we clearly need to do some relaxations of the minimization of @xmath60 with respect to ppt states .
let us define @xmath61 to be the set of quantum states supporting on @xmath48 , and denote the set of ppt states by @xmath62 .
we can first relax the minimization of @xmath60 to the smallest relative entropy distance between @xmath63 and the set @xmath62 .
see fig . 1 for the idea .
then applying some properties of quantum relative entropy , the problem can be relaxed to minimizing @xmath64 over all ppt states @xmath65 , where @xmath66 is the projection onto @xmath48 . noting that this is sdp computable
, we can further use sdp techniques to obtain the following bound @xmath67 interestingly , @xmath68 is additive under tensor product , i.e. , @xmath69 so we can overcome the difficulty of estimating the regularised relative entropy of entanglement .
the additivity of @xmath68 can be proved by utilizing the duality theory of sdp . a detailed proof of the additivity of @xmath68 is presented in the appendix .
more details about the applications of sdp in quantum information theory can be found in ref .
@xcite . to the state
@xmath65 taken from the set of ppt states @xmath62 .
assume that @xmath70 and @xmath71 give the smallest quantum relative entropy from @xmath63 to @xmath62 .
it is clear that @xmath72 and we show @xmath73 in proposition [ e eta bound er ] , where @xmath66 is the projection onto the support of @xmath1 .
this lower bound @xmath74 is powerful since it still works in the asymptotic setting due to its additivity under tensor product .
, scaledwidth=40.0% ] [ e eta bound er ] for any state @xmath1 , @xmath75 firstly , let us introduce a cptp map by @xmath76 . then for @xmath77 and @xmath78 , we have that @xmath79 where the first inequality is from the monotonicity of quantum relative entropy @xcite and the second inequality is due to the non - negativity of quantum relative entropy .
therefore , @xmath80 secondly , utilizing the weak duality theory of sdp , we can easily see that @xmath81 thus , @xmath82 finally , noting that @xmath74 is additive , we have that @xmath83 * ppt - assisted entanglement cost of @xmath55 : * applying the lower bound @xmath74 , we can now prove that the ppt - assisted entanglement cost of @xmath55 is still one ebit .
[ ppt cost ] for the state @xmath55 , @xmath84 firstly , suppose that @xmath85 and we can prove that @xmath86 by choosing a ppt state @xmath87 such that @xmath88 . secondly , we are going to prove @xmath89 . to see this ,
suppose that @xmath90.\end{aligned}\ ] ] noting that @xmath91 it is clear that @xmath92 .
moreover , @xmath93 which means that @xmath94 .
then @xmath95 is a feasible solution to the sdp ( [ e_eta ] ) of @xmath96 .
thus , @xmath97 and we can conclude that @xmath98 finally , noting that @xmath99 ( from ref .
@xcite , or simply applying the fact that any state in the support of @xmath55 is a @xmath2 maximally entangled state ) and @xmath100 , we have that @xmath101 .
our approach to evaluate the ppt - assisted entanglement cost is to combine the lower bound @xmath102 and upper bound @xmath35 .
this result provides a new proof of the entanglement cost of the rank - two @xmath0 antisymmetric state in ref .
@xcite .
moreover , our result gives a stronger argument that the entanglement cost under ppt operations of this state is still one ebit .
* ppt - assisted distillable entanglement of @xmath55 : * we can evaluate the ppt - assisted distillable entanglement of @xmath55 by the upper bound of rains bound and the sdp characterization of the one - copy ppt - assisted deterministic distillable entanglement @xcite . [
ppt distill ] for the state @xmath55 , @xmath103 on one hand , it is easy to calculate that @xmath104 .
then , @xmath105 on the other hand , suppose that @xmath106 with @xmath107 and @xmath108 .
it is easy to check that @xmath109 , which means that @xmath110 is a feasible solution to sdp ( [ prime ppt one - copy ] ) of @xmath111 .
therefore , @xmath112 finally , combining eq .
( [ ed upper ] ) and eq .
( [ ed lower ] ) , we have that @xmath113 . *
general irreversibility under ppt operations : * we have shown the irreversibility of the asymptotic entanglement manipulation of @xmath55 under ppt operations .
one can use similar technique to prove the irreversibility for any @xmath1 with spectral decomposition @xmath114 where @xmath115 interestingly , it holds that @xmath116 .
( see appendix ) .
more generally , we can provide a sufficient condition for the irreversibility under ppt operations and construct a general class of such states .
[ sdp condition ] if @xmath117 , then @xmath118 this is clear to see from @xmath119 indeed , we can obtain a more specific condition if we use logarithmic negativity @xmath120 instead of @xmath121 .
that is , for a bipartite state @xmath1 , if there exists a hermitian matrix @xmath122 such that @xmath123 and @xmath124 , then @xmath125 .
we further show the irreversibility in asymptotic manipulations of entanglement under ppt operations by a class of @xmath52 states in defined by @xmath126 where @xmath127 and @xmath128 with @xmath129 .
then the projection onto the range of @xmath130 is @xmath131 .
one can easily calculate that @xmath132 we then construct a feasible solution to the dual sdp ( [ e_eta ] ) of @xmath133 , i.e. , @xmath134 .
it can be checked that @xmath135 and @xmath136 .
thus , @xmath137 when @xmath138 , it is easy to check that @xmath139 .
therefore , @xmath140 the irreversibility under ppt operations is presented in fig . 2 for the state @xmath130 with @xmath129 .
exact values of @xmath141 and @xmath142 can be calculated by cvx @xcite assisted with qetlab @xcite . while the dash line depicts @xmath142 .
note that @xmath121 is the upper bound for ppt distillable entanglement while @xmath143 is the lower bound for ppt entanglement cost .
when @xmath144 , the asymptotic entanglement distillation under ppt operations is irreversible , i.e. , @xmath145 .
, scaledwidth=34.0% ] from theorem [ sdp condition ] , we further consider the existence of entanglement irreversibility in a given subspace . to be specific
, we derive a condition for the existence of irreversibility in a subspace @xmath146 with support projection @xmath47 by minimizing @xmath147 over any @xmath148 : @xmath149 for a bipartite subspace @xmath146 with support projection @xmath47 , if @xmath150 , then there exists a state @xmath148 such that @xmath125 .
* discussions * we prove that distillable entanglement can be strictly smaller than the entanglement cost under ppt operations , which implies the irreversibility in asymptotic manipulation of entanglement under ppt operations .
in particular , we prove that the ppt - assisted distillable entanglement of any rank - two @xmath0 antisymmetric state is strictly smaller than its ppt - assisted entanglement cost .
a byproduct is that there is a gap between the regularized rains bound and the asymptotic relative entropy of entanglement with respect to ppt states .
consequently , there is no unique entanglement measure in general for the asymptotic entanglement manipulation under ppt operations .
finally , we give an efficiently computable sufficient condition for the irreversibility under ppt operations .
however , @xmath143 is not tight for the @xmath0 anti - symmetric state @xmath151 , i.e. , @xmath152 @xcite .
how to improve the bound @xmath143 remains an interesting open problem .
rd would like to thank andreas winter for many inspirational discussions on the potential gap between the regularized rains bound and the asymptotic relative entropy of entanglement .
this work was partly supported by the australian research council under grant nos .
dp120103776 and ft120100449 . 1 c. h. bennett , d. p. divincenzo , j. a. smolin , and w. k. wootters , _ phys .
rev .
a _ * 54 * , 3824 ( 1996 ) . c. h. bennett , h. j. bernstein , s. popescu , and b. schumacher , _ phys . rev .
a _ * 53 * , 2046 ( 1996 ) .
g. vidal and j. i. cirac , _ phys .
rev .
lett . _ * 86 * , 5803 ( 2001 ) . g. vidal and j. i. cirac , _ phys .
rev . a _ * 65 * , 12323 ( 2001 ) .
g. vidal , w. dr , and j. i. cirac , _ phys .
rev .
lett . _ * 89 * , 27901 ( 2002 ) .
k. g. h. vollbrecht , r. f. werner , and m. m. wolf , _ phys .
rev .
a _ * 69 * , 62304 ( 2004 ) . m. f. cornelio , m. c. de oliveira , and f. f. fanchini , _ phys .
rev .
lett . _ * 107 * , 20502 ( 2011 ) .
m. horodecki , p. horodecki , and r. horodecki , _ phys .
rev .
lett . _ * 80 * , 5239 ( 1998 ) .
d. yang , m. horodecki , r. horodecki , and b. synak - radtke , _ phys .
rev .
lett . _ * 95 * , 190501 ( 2005 ) .
e. m. rains , _ ieee trans .
inf . theory _ * 47 * , 2921 ( 2001 ) .
t. eggeling , k. g. h. vollbrecht , r. f. werner , and m. m. wolf , _ phys .
rev .
lett . _ * 87 * , 257902 ( 2001 ) .
k. audenaert , m. b. plenio , and j. eisert , _ phys .
rev .
lett . _ * 90 * , 27901 ( 2003 ) .
f.g.s.l .
brando and m. b. plenio , _ nat .
phys .
_ * 4 * , 873 ( 2008 ) .
f.g.s.l .
brando and m. b. plenio , _ commun .
math .
phys . _ * 295 * , 829 ( 2010 ) .
f.g.s.l .
brando and g. gour , _ phys .
rev .
lett . _ * 115 * , 70503 ( 2015 ) .
o. krueger and r. f. werner , arxiv prepr .
quant - ph/0504166 ( 2005 ) .
see also url : http://qig.itp.uni-hannover.de/qiproblems/20 .
m. b. plenio and s. virmani , _ quantum inf .
comput . _ * 7 * , 1 ( 2007 ) .
l. vandenberghe and s. boyd , _ siam rev . _
* 38 * , 49 ( 1996 ) . c. h. bennett , d. p. divincenzo , c. a. fuchs , t. mor , e. rains , p. w. shor , j. a. smolin , and w. k. wootters , _ phys .
rev . a _ * 59 * , 1070 ( 1999 )
. p. m. hayden , m. horodecki , and b. m. terhal , _
j. phys .
a. math .
gen . _ * 34 * , 6891 ( 2001 ) .
k. audenaert , b. de moor , k. g. h. vollbrecht , and r. f. werner , _ phys .
rev .
a _ * 66 * , 32310 ( 2002 ) .
x. wang and r. duan , arxiv:1605.00348 ( 2016 ) .
v. vedral , m. b. plenio , m. a. rippin , and p. l. knight , _ phys .
rev .
lett .
_ * 78 * , 2275 ( 1997 )
. v. vedral and m. b. plenio , _ phys .
rev .
a _ * 57 * , 1619 ( 1998 ) .
v. vedral , m. b. plenio , k. jacobs , and p. l. knight , _ phys .
rev . a _ * 56 * , 4452 ( 1997 ) .
g. vidal and r. f. werner , _ phys .
rev . a _ * 65 * , 32314 ( 2002 )
. m. b. plenio , _ phys .
rev .
lett . _ * 95 * , 90503 ( 2005 ) .
x. wang and r. duan , arxiv:1601.07940 ( 2016 ) .
m. hayashi , quantum information ( springer , 2006 ) .
e. chitambar and r. duan , _ phys .
rev .
lett . _ * 103 * , 110502 ( 2009 ) .
j. watrous , theory of quantum information , university of waterloo ( 2011 ) .
g. lindblad , _ commun .
math .
phys .
_ * 40 * , 147 ( 1975 ) .
a. uhlmann , _ commun .
math .
phys . _ * 54 * , 21 ( 1977 ) .
f. yura , _ j. phys . a math .
gen . _ * 36 * , 15 ( 2003 ) .
m. grant and s. boyd .
cvx : matlab software for disciplined convex programming , http://cvxr.com ( 2014 ) .
m. c. grant and s. p. boyd , in _ recent adv .
learn .
control _
( springer , 2008 ) , pp . 95110 .
nathaniel johnston , , http://qetlab.com ( 2015 ) . |
as asymptotic giant branch ( agb ) stars evolve toward compact objects with surrounding planetary nebulae ( pne ) , mass - loss occurs in the spherically symmetric circumstellar envelopes ( cses ) with low ( @xmath930kms@xmath6 ) expansion velocities . despite the spherical symmetry in the progenitor objects , pne typically exhibit morphologies , such as axisymmetric structures , that are clearly not formed through normal agb mass - loss mechanisms @xcite .
maser emission ( sio , h@xmath1o and oh ) from agb stars can provide insights into the structure and dynamics of the cse . in late - type stars ,
the conditions for the formation of the various species of masers are typically met in a progression of zones .
the sio masers occur relatively close to the star and have velocities near the radial velocity of the star .
the h@xmath1o and oh masers are found in the circumstellar wind region at progressively further distances from the star .
the h@xmath1o and oh spectral profiles are often double peaked , and the velocity range covered by the oh is larger than that of the h@xmath1o .
there is a small but growing number of evolved stars which have been shown to deviate from this prototypical agb envelope .
these objects have been dubbed
water - fountain " sources due to their high - velocity h@xmath1o masers . to date
there are three confirmed water - fountain sources : iras16342@xmath03814 @xcite , iras19134 + 2131 @xcite , and w43a @xcite .
the h@xmath1o masers associated with these objects are characterized by wide velocity ranges @xmath10kms@xmath6 .
specifically , for iras16342@xmath03814 , iras19134 + 2131 , and w43a , @xmath11 is 259kms@xmath6 , 132kms@xmath6 and 180kms@xmath6 respectively ( * ? ? ?
* and references therein ) .
these velocties are much greater than those of the typical oh / ir star in which differences between the blue- and red - shifted peaks of the oh masers are @xmath22040kms@xmath6 @xcite and roughly 75% of this range ( 1530 kms@xmath6 ) for the h@xmath1o masers @xcite .
in fact , for the two sources ( iras16342@xmath03814 and w43a ) that have detected oh masers , the h@xmath1o masers have a larger @xmath11 than the corresponding oh . in addition to the unusual spectral characteristics , the h@xmath1o masers associated with the water - fountain sources appear to exhibit unique spatial structures as revealed by radio interferometric images . in the first very long baseline interferometry ( vlbi )
study of the h@xmath1o masers toward w43a , @xcite showed that the water masers are formed in a collimated precessing jet with a true 3-dimensional space velocity of 145kms@xmath6 .
studies of iras16342@xmath03814 @xcite and iras19134 + 2131 @xcite have yielded similar bipolar maser distributions albeit without the precession observed in w43a .
the h@xmath1o masers toward iras16342@xmath03814 and w43a are also more extended than the oh masers in these sources .
the bipolar jets traced by the h@xmath1o masers , presumably along the polar axis of the star , may represent the onset of the axisymmetric morphologies that typify pne .
the dynamical ages of the jets for iras19134 + 2131 and w43a are estimated to be @xmath250yr and @xmath240yr respectively @xcite . that for iras16342 - 3814 is estimated to be @xmath2150yr @xcite .
thus the evolutionary stage that these water - fountain sources represent is likely very short , so such objects must be quite rare .
maser emission from [email protected] was first observed by @xcite who classified it as a type ii oh / ir star because of its characteristic double - peaked 1612-mhz oh maser profile .
the star has been associated with the infrared source iras18139 - 1816 @xcite which is @xmath226arcsec away .
the exceptional spectral characteristics of the h@xmath1o masers toward [email protected] , were first highlighted by @xcite .
@xcite used the very large array ( vla ) to show that the anomalous " h@xmath1o and oh maser emission was coincident to within 1@xmath12 , and therefore belonged to the same source .
@xcite also found double - peaked profiles for both the oh and the h@xmath1o with the peaks of the oh separated by @xmath223kms@xmath6 and h@xmath1o peaks separated by nearly twice this amount , @xmath242kms@xmath6 .
although the velocity range for the h@xmath1o is not as wide as the other water - fountain sources , @xcite found that the shape and variations in the spectra are consistent with the water - fountain class and are compatible with an axisymmetric wind .
engels suggested that [email protected] may represent an earlier evolutionary stage of water fountain , and that the h@xmath1o maser velocities may increase with increasing post - agb wind speeds .
although the spectral characteristics provide a hint that [email protected] is related to the water - fountain sources , to verify that it is truly a member of this class requires high - resolution vlbi imaging of the h@xmath1o maser emission .
we therefore observed the h@xmath1o masers associated with [email protected] using the very long baseline array ( vlba ) in order to map the maser distribution . in this article , we present the first vlbi images of the h@xmath1o masers toward [email protected] and discuss the nature of this interesting source .
we observed the 22.2 ghz h@xmath1o and both the 1612 mhz satellite - line and the 1667 mhz main - line oh maser emission from [email protected] ( @xmath13 , j2000 ) .
observations of the h@xmath1o masers occurred on 2004 june 21 for 5 hrs starting at 04:43 ut ( jd 2453177.7 ) .
observations of the oh masers occurred in two 5hr epochs on 2004 march 16 starting at 11:04 ut ( jd 2453081.0 ) and 2004 july 25 starting at 02:29 ut ( jd 2453211.6 ) . [email protected] and three continuum calibrators ( 1751 + 096 , j1832@xmath02039 , and 3c345 ) were observed using the 10 stations of the vlba .
the vlba is operated by the national radio astronomy observatory ( nrao ) .
reference frequencies of 22.23508 , 1.61223 , and 1.66736ghz were used for the h@xmath1o , the oh satellite - line , and the oh main - line maser transitions respectively .
the h@xmath1o data were recorded in dual circular polarization using two 8-mhz ( 112.6kms@xmath6 ) bands centered on the local standard of rest ( lsr ) velocity of @xmath058.0kms@xmath6 .
system temperatures and point source sensitivities were on the order of @xmath2120k and @xmath29jyk@xmath6 respectively .
the oh data were recorded in dual circular polarization using four 0.5-mhz ( 92.7kms@xmath6 ) bands centered on the lsr velocity of @xmath058.0kms@xmath6 .
system temperatures and point source sensitivities were on the order of @xmath2160k and @xmath29.5jyk@xmath6 respectively .
the data were correlated at the vlba correlator operated by nrao in socorro , new mexico . auto and cross - correlation spectra consisting of 512 channels with channel spacings of 15.63khz ( @xmath20.22kms@xmath6 ) and 0.98khz ( @xmath20.18kms@xmath6 ) for the h@xmath1o and oh respectively were produced .
calibration was performed in accordance with standard vlba spectral - line procedures using the astronomical image processing system ( aips ) maintained by nrao .
the bandpass response was determined from scans on the continuum calibrators and was used to correct the target source data .
a preliminary inspection of the auto - correlation ( ac ) spectra for the 10 antennas at this point showed that the 1667 mhz main - line oh transition was not detected toward [email protected] .
both the h@xmath1o and the 1612 mhz satellite line of oh were detected in the ac spectra of all antennas .
the time - dependent gains of all antennas relative to a reference antenna were determined by fitting a template ac spectrum ( from the reference antenna with the target source at a high elevation ) to the ac spectra of each antenna .
the absolute flux density scale was established by scaling these gains by the system temperature and gain of the reference antenna .
errors in the gain and pointing of the reference antenna and the atmospheric opacity contribute to the error in the absolute amplitude calibration , which is accurate to about 1520% . to correct any instrumental delay , a fringe fit was performed on the continuum calibrator scans and residual group delays for each antenna were determined .
residual fringe - rates were obtained by fringe - fitting a strong reference feature in the spectrum of each maser transition . at this stage in the reduction ,
no fringes for the 1612 mhz oh line were detected on baselines other than the shortest ( los alamos pie town ) vlba baseline . the likely cause for this non - detection is localized interstellar scattering along the line - of - sight to [email protected] , which is relatively close to the galactic center , a direction known to exhibit anomalously high scattering on certain lines of sight @xcite .
thus , the remaining discussion applies only to the h@xmath1o maser data .
the fringe - rate solutions from the fringe fit were applied to all channels in the h@xmath1o maser spectrum . an iterative self - calibration and imaging procedure
was then performed on the reference channel .
the resulting residual phase and amplitude corrections were applied to all channels in the band .
a low - resolution cube of images @xmath14 pixels ( @xmath2@xmath15mas ) was formed covering the observed velocity range to locate regions of emission using a synthesized beam of @xmath16mas .
two regions of emission were identified corresponding to spectral channel ranges from @xmath037.8kms@xmath6 to @xmath029.4kms@xmath6 and from @xmath086.2kms@xmath6 to @xmath075.7 kms@xmath6 .
full - resolution cubes of images @xmath17 pixels ( @xmath2@xmath18mas ) were generated for each range of channels using a synthesized beam of @xmath19mas .
the rms off - source noise ( @xmath20 ) was typically 3mjybeam@xmath6 in all channels . in order to extract maser component parameters , two - dimensional gaussian functions were fit to the emission in the individual channel maps .
image quality was assessed using the off - source rms noise and the deepest negative pixel in the image .
a cutoff flux density was conservatively set at 10@xmath20 .
features with flux densities greater than this cutoff were fit with gaussians to determine component parameters .
errors in the fitted positions of identified features were computed following the methods outlined in @xcite .
these errors ranged from 5@xmath21as for features with high signal - to - noise , to 60
@xmath21as for features with lower signal - to - noise .
because the emission from each maser feature extends over multiple spectral channels , it is desirable to determine a single velocity , flux density and position for each maser feature . in lieu of a full three - dimensional ( 3-d ) gaussian fit to the image cube
, we used a flux - density - squared weighted averaging scheme to calculate nominal positions and velocities for each maser from the previously identified features .
this method takes into account the increased significance of stronger emission channels .
the velocities and relative offset positions thus represent weighted averages over the number of spanned channels . the assigned flux density is simply the peak value over the same range of channels .
figure [ spot_map ] shows the spectral ( upper sub - panels ) and spatial ( lower sub - panels ) distributions of the h@xmath1o masers toward [email protected] from the analysis of our vlba images .
panel ( a ) shows the entire range of h@xmath1o maser emission from the star .
panels ( b ) and ( c ) show enlarged views of the blue - shifted masers to the north and the red - shifted masers to the south respectively . focusing on the upper sub - panel of figure [ spot_map](a ) ,
we see the previously observed spectral characteristics of the [email protected] .
the 1612 mhz oh masers form a double - peaked profile with peaks at @xmath068.0 and @xmath043.7kms@xmath6 . the center of the profile is roughly @xmath055.8kms@xmath6 , which is consistent with the stellar radial velocity of @xmath22kms@xmath6 previously determined from the oh emission @xcite .
the h@xmath1o spectrum also has a double - peaked profile , however , the h@xmath1o masers have a greater velocity extent with peaks at @xmath081.7 and @xmath033.3kms@xmath6 . the midpoint between these peaks ( @xmath057.5kms@xmath6 )
is slightly offset from the stellar velocity .
the lower sub - panel of figure [ spot_map](a ) shows that the masers occupy two distinct regions oriented roughly north
south on the sky with an angular separation of @xmath2109mas .
the position angle of the axis of separation between the centers of the northern and southern maser regions is 1.5@xmath3 east of north .
this double - lobed distribution is consistent with vlbi observations of other water - fountain sources that show the masers occupy widely separated clumps along a presumed polar axis .
separations of @xmath23000mas ( 6000 au ) , @xmath2150mas ( 2400 au ) and @xmath2700mas ( 1700 au ) were determined for iras16342@xmath03814 , iras19134 + 2131 and w43a respectively @xcite ( note that the distances to all sources are somewhat uncertain , with w43a the most reliably known ) . the linear separation of the two maser regions toward [email protected] is approximately @xmath23 , where @xmath24 is the distance in kpc .
unfortunately , the distance to [email protected] is unknown .
if we assume the star is associated with the galactic center @xcite , at @xmath25kpc , then the linear separation between the blue- and red - shifted masers is @xmath2870au . upon closer inspection of the blue- and red - shifted maser regions ( figure [ spot_map ] ( b ) and ( c ) ) , we find that the masers are aligned along arc - like structures .
the blue- and red - shifted arcs are approximately 10 and 12mas ( 80 and 96au at 8kpc ) across respectively , corresponding to an opening angle for the jet that likely drives their formation of 1013@xmath3 .
these arcs have an orientation that is roughly perpendicular to the separation ( polar ) axis .
one might expect such a morphology in the case of an edge - on shock front propagating outward into a slower ambient medium .
such arc - like structures have also been observed for the h@xmath1o masers toward iras 16342@xmath03814 , and have been shown to have a global outward motion @xcite .
further observations of [email protected] will be required to measure the proper motions of its h@xmath1o masers .
although the radial velocity range for the h@xmath1o is not as wide as the more well - known water - fountain sources , @xcite found that the shape and variations in the h@xmath1o maser spectra are consistent with this class .
the velocity difference between the blue- and red - shifted h@xmath1o masers also appears to be increasing over time from early values of @xmath26kms@xmath6 @xcite to our most recent measurement of @xmath27kms@xmath6 .
values of @xmath11 based on the most dominant features in the h@xmath1o spectra as published in the literature are listed in table [ accel_table ] and plotted in figure [ accel_plot ] .
all values prior to those in this work are determined from low spatial resolution ( i.e. single dish and vla ) observations and are therefore subject to spatial blending of multiple maser features in a single spectral channel . from figure [ accel_plot ]
we see that @xmath11 is obviously not constant over time , and that the masers are undergoing some form of acceleration ( possibly non - linear ) .
since the errors in @xmath11 as determined from the literature are not well known , we chose to perform a simple least - squares fit to the points assuming constant acceleration .
this fit yields a value of @xmath28kms@xmath29 or @xmath30kms@xmath6yr@xmath6 for the relative acceleration of the blue- and red - shifted masers in the radial direction .
the 3-d acceleration is likely larger by a factor of @xmath31 , where @xmath32 is the inclination of the jet .
the true 3-d velocities and accelerations of the masers remain to be determined through future vlbi proper motion studies . as mentioned previously , the distance to [email protected] is unknown , which makes estimating the dynamical age of the outflow difficult . if we make several assumptions , namely : that the motions of the masers in the plane of the sky are comparable to the line - of - sight acceleration , that the dynamical center of the outflow is the midpoint along the axis of separation between the two maser regions , that the distance to the source is 8kpc , and that the masers have zero initial velocity , then an upper limit to the dynamical age of the outflow can be computed . under these assumptions , we compute a dynamical age of @xmath2109yr using the current spatial extent of the masers and the previously determined acceleration .
similarly , a dynamical age of @xmath271yr is determined using the current expansion velocity of 24.2kms@xmath6 and the above acceleration .
these two ages are roughly consistent with each other and with the ages computed for the three other water - fountain sources . for comparison , the dynamical ages of iras16342@xmath03814 , iras19134 + 2131 and w43a have been estimated to be @xmath2150yr @xcite , @xmath250yr @xcite and @xmath240yr @xcite respectively .
it is interesting to note that while the small velocity spread between the blue- and red - shifted masers of [email protected] relative to the other water - fountain sources would seem to indicate that the outflow is relatively young , the upper limit on the dynamical age is intermediate to those of the other sources in the class . assuming the acceleration remains constant at 0.68kms@xmath6yr@xmath6 , then the time to reach an outflow velocity comparable to the other water fountain sources ( @xmath33kms@xmath6 ) is only about 220yr or roughly twice the current dynamical age of [email protected] .
using the vlba we have produced the first high - resolution interferometric images of the h@xmath1o masers toward the oh / ir star [email protected] .
these images show that the masers lie in a double - lobed structure typical of objects belonging to the water - fountain class of sources .
we also find that the two groupings of masers are arranged in arc - like structures suggestive of a bow - shock morphology .
the spatio - kinematic structure determined from our vlba observations provides strong evidence that [email protected] is indeed a water fountain source , and that the outflow is accelerating in the line - of - sight direction . in future observations
, we hope to map the scatter - broadened oh masers using a shorter - baseline connected element interferometer such as the vla or the multi - element radio linked interferometer ( merlin ) .
such observations should verify that the h@xmath1o masers lie further from the underlying star than the oh masers ; a characteristic evident in the other water - fountain sources .
we also plan to undertake a multi - epoch vlbi campaign to study the proper motions of the h@xmath1o masers .
such observations should allow us to characterize the three - dimensional motion of the masers , determine the distance to [email protected] , and determine a dynamical center and age for the outflow .
cccl 1985.30 & 34.7 & 0.2 & ( 1 ) , strongest features - matched to hi - res ep .
+ 1985.58 & 36.1 & 0.2 & ( 1 ) , strongest features - matched to hi - res ep .
+ 1987.45 & 38.4 & 0.16 & ( 2 ) , feature
f - feature n + 1988.51 & 38.9 & 0.16 & ( 2 ) , feature f - feature n + 1992.47 & 42.3 & 2.6 & ( 3 ) , ( blend of blue and red features ) + 1995.19 & 44.5 & 0.16 & ( 2 ) , feature b - feature o + 1999.09 & 45.5 & 0.16 & ( 2 ) , feature b - feature o + 2004.57 & 48.4 & 0.2 & ( 4 ) , strongest features + | we present observational evidence that the oh / ir star [email protected] is the fourth in a class of objects previously dubbed water - fountain " sources . using the very long baseline array
, we produced the first images of the h@xmath1o maser emission associated with [email protected] .
we find that the masers are located in two compact regions with an angular separation of @xmath2109mas on the sky .
the axis of separation between the two maser regions is at a position angle of 1.5@xmath3 east of north with the blue - shifted ( @xmath4 to @xmath5 kms@xmath6 ) masers located to the north and the red - shifted ( @xmath7 to @xmath8kms@xmath6 ) masers to the south .
in addition , we find that the blue- and red - shifted masers are distributed along arc - like structures @xmath21012mas across oriented roughly perpendicular to the separation axis .
the morphology exhibited by the h@xmath1o masers is suggestive of an axisymmetric wind with the masers tracing bow shocks formed as the wind impacts the ambient medium .
this bipolar jet - like structure is typical of the three other confirmed water - fountain sources . when combined with the previously observed spectral characteristics of [email protected] , the observed spatio - kinematic structure of the h@xmath1o masers provides strong evidence that [email protected] is indeed a member of the water - fountain class . |
99 see for example , nakano t 2004 , plenary talk at nstar 2004 , grenoble , march 2004 diakonov d , petrov v and polyakov m 1997 _ z. phys .
a_*359 * 305 alt c _ et al .
_ 2004 _ phys .
lett . _ * 92 * 042003 arndt r a , briscoe w j , strakovsky i i , workman r l , and pavan m m 2004 _ phys . rev .
c _ * 69 * 035213 diakonov d and petrov v , 2004 _ phys .
d _ * 69 * 094011 eidelman s _
_ 2004 _ phys .
b _ * 592 * 1 koch r 1985 _ z. phys .
c _ * 29 * , 597 ; hhler g 1983 _ pion nucleon scattering _
, landoldt brnstein vol . *
i/9b2 * , edited by schopper h ( springer verlag ) .
cutkosky r e _ et al .
_ 1980 in _ proceedings of 4th international conference on baryon resonances , toronto , canada , july 1980 _ , published in baryon 1980:19 ( qcd161:c45:1980 ) ; cutkosky e e and wang s 1990 _ phys .
* 42 * 235 manley d m and saleski e m 1992 _ phys .
d _ * 45 * 4002 azimov ya i , arndt r a , strakovsky i i , and workman r l 2003 _ phys .
c _ * 68 * 045204 arndt r a , azimov ya i , polyakov m v , strakovsky i i , and workman r l 2004 _ phys . rev .
c _ * 69 * 035208 arndt r a , strakovsky i i , and workman r l 2003 _ phys .
c _ * 68 * 042201 kabana s , hep - ex/0406032 kuznetsov v , hep - ex/0409032 bai j z _ et al . _ 2004 _ phys .
d _ * 70 * 012004 azimov ya i and strakovsky i i 2004 _ phys .
c _ * 70 * 035210 dzierba a _ et al .
_ 2004 _ phys .
d _ * 69 * 051901 stepanyan s _ et al .
_ 2003 _ phys .
lett . _ * 91 * 252001 hicks k , burkert v , kudryavtsev a , strakovsky i , and stepanyan s , submitted to _ phys . rev .
d _ hep - ph/0411265 karliner m and lipkin h 2004 _ phys . lett .
b _ * 597 * 309 polyakov m v and rathke a 2003 _ eur .
j. a _ * 18 * 691 gusey v and polyakov m v , hep - ph/0501010 cccc collaboration & mass ( mev ) & width ( mev ) & ref + dpp & 1710 ( input ) & @xmath4840 & @xcite + kh & 1723@xmath499 & 120@xmath4915 & @xcite + cmu & 1700@xmath4950 & 90@xmath4930 & @xcite + ksu & 1717@xmath4928 & 480@xmath49230&@xcite + gwu & 1710 & no state ! & @xcite + | given the existing empirical information about the exotic @xmath0 baryon , we analyze possible properties of its @xmath1-partners , paying special attention to the nonstrange member of the antidecuplet @xmath2 .
the modified @xmath3 partial - wave analysis presents two candidate masses , 1680 mev and 1730 mev . in both cases
, the @xmath2 should be rather narrow and highly inelastic .
our results suggest several directions for experimental studies that may clarify properties of the antidecuplet baryons , and structure of their mixing with other baryons .
recent experimental evidence from the graal and star collaborations could be interpreted as observations of a candidate for the @xmath0 nonstrange partner . .
results from a wide range of recent experiments are consistent with the existence of an exotic s=+1 resonance , the @xmath4 with a narrow width and a mass near 1540 mev @xcite .
direct width determinations have been hindered by the limitations of experimental resolution , resulting in upper bounds of order 10 mev .
the quantum numbers of this state remain unknown , though a prediction of @xmath5 = @xmath6 was obtained in the work @xcite that provided motivation for the original search .
additional information related to the assignment of unitary partners is due to a recent experimental result @xcite giving evidence for one further explicitly exotic particle @xmath7 , with a mass @xmath8 mev and width @xmath9 mev ( _ i.e. _ less than resolution ) .
such a particle had been expected to exist as a member of an antidecuplet , together with the @xmath0 .
however , the soliton calculation of the mass difference requires some assumptions .
in particular , it depends on the value of the @xmath10-term , which is the subject of controversy .
its value , taken according to the latest @xmath3 data analysis @xcite , leads to an antidecuplet mass difference of about 110 mev , instead of the originally predicted 180 mev @xcite .
so , if the states @xmath11 @xcite and @xmath12 are indeed members of the same antidecuplet , then , according to the gell - mann - okubo rule , the mass difference of any two neighboring isospin multiplets in the antidecuplet should be constant and equal @xmath13 which fits the gw said @xmath10-term result very well @xcite .
this shift also effects the masses of other unitary partners of the @xmath0 : nucleon - like and @xmath14-like .
the supposed antidecuplet , with @xmath14- and @xmath15-masses determined by the gell - mann
okubo rule , looks today as shown on fig .
[ fig : g1 ] . due to @xmath1-violating mixing with lower - lying nucleon - like octet states ,
@xmath16 may shift upward , and reach about 1680 mev @xcite . mixing with higher - lying nucleon - like members of exotic 27- and 35-plets
may also play a role .
the state @xmath17 , though listed in the pdg baryon summary table @xcite as a 3-star resonance and used as input in the @xmath0 prediction @xcite , is not seen in the latest analysis of pion - nucleon elastic scattering data ( see table [ tbl : tbl1 ] ) .
studies which have claimed to see this state have given widely varying estimates of its mass and width ( from @xmath18 mev to @xmath19 mev for the mass and from @xmath20 mev to @xmath21 mev for the width ) .
branching ratios have also been given with large uncertainties ( 1020% for @xmath22 , 4090% for @xmath23 , and so on ) , apart from one which has been presented with much greater precision ( @xmath24 for @xmath25 ) .
in any case , the pdg width of @xmath17 seems to be too large for the partner of the narrow @xmath0 .
it would be more natural for members of the same unitary multiplet to have comparable widths . as has been suggested recently ( see refs .
@xcite ) , any standard pwa by itself tends to miss narrow resonances .
for this reason , we considered @xcite a modified pwa , assuming the existence of a narrow resonance , and compared the quality of fit with and without such structures ( more detailed description of formulas see in ref .
@xcite ) .
such an approach was used initially to look for light nucleon resonances @xcite .
this method , applied to studies of the @xmath4 @xcite , though unable to determine the @xmath0 quantum numbers , placed a tight limit on the width , in full agreement with the results of other approaches .
we subsequently used this method @xcite to search @xmath3 scattering data for a narrow nucleon - like state assumed to be a member of the antidecuplet , accompanying the @xmath4 .
the two candidate masses , m@xmath26 = 1680 mev and 1730 mev , would necessary be quite inelastic with @xmath27 0.5 mev and 0.3 mev , respectively .
some support for a narrow structure in this mass region has recently been obtained in a preliminary report based on direct measurements by the star @xcite and graal @xcite collaborations .
it should be noted that there have been other publications finding no evidence for the existance of the @xmath0 .
these findings need more detailed consideration .
though the present non - observation data require some suppression of exotic production , as compared to conventional hadrons , they do not definitively exclude the existence of the @xmath0 and/or its companions / analogs .
for example , analysis of the bes data @xcite in ref .
@xcite shows that indeed there is some suppression of @xmath28-production .
however , suppression of the exotics at this level could have a quite natural explanation .
there is also a statement that the observed peak of @xmath0 could be due to a kinematical reflection of some of known resonances .
a particular consideration has been suggested by dzierba _
et al . _
@xcite addressed to the clas analysis @xcite .
the specific model used by dzierba _
et al .
_ has , however , been criticized @xcite , and may not be a serious concern for the clas results .
we should emphasize here that if the present evidence for the @xmath28 turns out to be incorrect , we would have to answer a different , but also difficult , question : why do we _ not _ see exotic hadrons ?
we take here , for the sake of argument , the position that the @xmath28 does exist , but its production is governed by different mechanisms . though we essentially agree with suggestions of karliner and lipkin @xcite about how to clarify the problem , we think that , first of all , it is especially important to reliably confirm the existence of the @xmath28 in the processes where it has been reported .
new data are being collected for this purpose , by several collaborations , and one would hope for a definitive answer within a year . that is why , at the moment , we will assume the @xmath0 ( as well as other multi - quark hadrons ) being existent , and will discuss some consequences of this fact ( for details , see ref .
@xcite ) . to summarize ,
given our current knowledge of the @xmath0 , the state commonly known as the @xmath17 is not the appropriate candidate to be a member of the antidecuplet together with the @xmath0 .
instead , we suggest candidates with nearby masses , @xmath29 ( more promising ) and/or @xmath30 ( less promising , but not excluded ) .
our analysis suggests that the appropriate state should be rather narrow and very inelastic .
similar considerations have been applied to the @xmath31 , assumed to be also a member of the same antidecuplet .
it should also be quite narrow .
how reliable are our theoretical predictions ?
they have , indeed , essential theoretical uncertainties .
we have yet to establish the existence of the ( narrow ) state originally associated with the @xmath17 .
moreover , we have assumed the presence of only one state with @xmath32 , either @xmath29 or @xmath30 . if both exist with the same spin and parity ,
our conclusions should be reconsidered .
futhermore , we use the mixing angle @xmath33 , taken from ref . @xcite , which was actually determined through formulas containing the @xmath10-term ( just as the mass difference in the antidecuplet ) .
if we use parameters corresponding to more recent information , for both the @xmath10-term and the mass difference , we obtain larger mixing , up to @xmath34 . with our formulas
, this would most strongly influence the partial width @xmath35 , increasing it to about 15 mev .
other partial widths of @xmath36 change not so dramatically , and the total width appears to remain not higher than @xmath37 mev
. such a width could well be measured , but not in elastic scattering , because of an expected very small elastic branching ratio .
note , however , that the above large value for @xmath38 may appear problematic , since the formulas of ref .
@xcite assume linearisation with respect to @xmath1-violation , and need to be reconsidered if the violation appears to be large .
nevertheless , even having in mind all theoretical uncertainties , we can suggest several directions for experimental studies .
first of all , one should search for possible new narrow nucleon state(s ) in the mass region near 1700 mev .
searches may use various initial states , ( _ e.g. _ , @xmath3 collision or photoproduction ) .
we expect the largest effect in the @xmath39 final state ( mainly through @xmath40 , though it is forbidden by @xmath1 ) . the final states @xmath41 and @xmath42 may also be interesting and useful , especially the ratio of @xmath41 and @xmath3 partial widths , as the latter is very sensitive to the structure of the octet antidecuplet mixing .
another interesting possibility to separate antidecuplet and octet components of @xmath2 is provided by comparison of photoexcitation amplitudes for neutral and charged isocomponents of this resonance , the point being that the antidecuplet component contribution to the photoexitation of the charged @xmath2 is strongly suppressed ( see details in ref .
@xcite ) .
on the other hand , such a relatively simple picture of mixing can not reproduce our small value(s ) of @xmath43 .
we assumed in our analysis that this could result from more complicated mixing with several other multiplets @xcite .
such a possibility was recently confurmed @xcite . for @xmath11
, attempts to measure the total width are necessary , though it could possibly be even smaller than @xmath44 . branching ratios for @xmath45 and @xmath46 , in relation to @xmath47 , are very interesting .
these may give important information on the mixing of antidecuplet baryons with octets and higher @xmath1-multiplets .
the work was partly supported by the u. s. department of energy grant de
fg0299er41110 , by the jefferson laboratory , by the southeastern universities research association under doe contract de ac0584er40150 , by the russian state grant ss1124.2003.2 . |
it is a common knowledge ( see , for example @xcite ) that the wave function @xmath5 of the @xmath0-spin particle , which can describe the state of a two - level atom and consists of two components @xmath6 and @xmath7 , can be represented in the form of a spinor @xmath8 in the case of the pure particle s state , the hermitian density matrix has the form of a @xmath9 matrix @xmath10 where the diagonal matrix elements satisfy the normalization condition @xmath11 .
any rotation in the three - dimensional space , determined by the euler angles @xmath12 varying in the domain @xmath4 , @xmath13 , @xmath14 , is associated with a @xmath9 unitary matrix @xmath15 & \sin \frac \theta 2\exp \left[-\frac{i\left ( \varphi -\psi \right ) } 2\right ] \\
-\sin \frac \theta 2\exp \left[\frac{i\left ( \varphi -\psi \right ) } 2\right ] & \cos \frac \theta 2\exp \left[-\frac{i\left ( \varphi + \psi \right ) } 2\right ] \end{array } \right ) .
\label{rotation}\ ] ] under rotation , the hermitian density matrix @xmath1 is transformed as @xmath16 where the @xmath9 unitary matrix @xmath17 is a representation of the angle that describes the matrix elements of the matrix @xmath17 . for the diagonal elements of the hermitian density matrix
@xmath18 we get @xmath19 here @xmath20 is the wigner d - function and @xmath21
. the diagonal elements of the density matrix of the quantum state take nonnegative values , and their sum is equal to unity .
the physical meaning of these elements is that they are a probability to measure the value of the spin projection on the fixed axis in a space .
then , we introduce the notation @xmath22 where the function @xmath23 is the marginal distribution , i.e. , the probability of finding the spin projection @xmath24 on the fixed axis in the space for the @xmath0-spin particle . from eq .
( [ den!mar ] ) , we see that the marginal distribution also depends on the euler angles @xmath25 as parameters .
this distribution function is normalized for all values of the euler angles . from the structure of eq .
( [ diagon!elem ] ) , it follows that the marginal distribution @xmath23 depends only on two euler angles , and does not depend on the angle @xmath26 of the rotation .
in fact , according to eq .
( [ diagon!elem ] ) , one can obtain the marginal distribution @xmath23 of the @xmath0-spin particle in the following form @xmath27 and @xmath28 hence , the family of the probability distribution functions of the @xmath0-spin projection is parametrized by the point s coordinates @xmath29 , @xmath30 on the sphere of unity radius .
this parametrization coincides with the physical meaning of the marginal distribution in the sense that the distribution function @xmath23 is the probability to observe the spin projection @xmath24 if we measure this spin projection on the quantization axis which is parallel to the vector normal to the surface of the sphere of the unity radius in the point with the coordinates @xmath29 and @xmath30 . if we know the positive , normalized marginal distribution @xmath23 , then , as it was shown in @xcite , the matrix elements @xmath31 can be calculated with the help of the measurable marginal distribution @xmath23 of the particle with an arbitrary spin @xmath32 and the values of indices @xmath33 by means of the relation @xmath34 where @xmath35 and the integration leads over the rotation angles @xmath25 @xmath36 in the case of the @xmath0-spin ( i.e. , @xmath37 ) , eq . ( [ fig1 ] )
reads @xmath38 where @xmath39 . calculating all non - zero integrals with d - functions and the marginal distributions and taking into account the expressions of 3j - symbols ( see , for example , @xcite ) , we can rewrite eq . ( [ always ] ) as follows @xmath40 where @xmath39 , and @xmath24 is the imaginary unity .
the substitution of @xmath41 in eq .
( [ reverse!mat!den ] ) leads to the following result : @xmath42 and in all the other cases @xmath43 then , we checked by the direct calculation that the right - hand side of eq .
( [ always ] ) is equal to the density matrix ( [ m!d ] ) .
one can conclude , that given a measurable marginal distribution of a particle , whose state is described in terms of spinor , one can reconstruct the state density matrix by means of eq .
( [ always ] ) .
the classical - like description of quantum mechanics in terms of the positive , normalized marginal distribution can be easily understood in the case of the @xmath0-spin particle .
it is a common knowledge @xcite , that the spin operator of the @xmath0-spin particle has the form @xmath45 where the pauli matrixes are @xmath46 the wave function of the @xmath0-spin particle can be represented in the form of spinor ( [ spinor ] ) .
when the direction of the particle s spin coincided with the positive direction of the @xmath47 axis in the coordinate system ( @xmath47 , @xmath48 , @xmath49 ) , the wave function of spinor has the form @xmath50 and in the case of a contrary direction of the spin with respect to the @xmath47 axis , the wave function reads @xmath51 analogously for the @xmath48-projection , when the direction of the particle s spin is @xmath52 or @xmath53 on the @xmath48 and @xmath49 axes , we have @xmath54 @xmath55 and for the @xmath49-projection @xmath56 @xmath57 for all values of the @xmath0-spin directions on the @xmath47 , @xmath48 and @xmath49 axes the density matrices have the following forms @xmath58 @xmath59 and @xmath60 applying eqs .
( [ mar!dist!spinor!pos ] ) and ( [ mar!dist!spinor!op ] ) , we obtain the marginal distributions for the different values of the @xmath0-spin directions on the @xmath47 , @xmath48 and @xmath49 axes @xmath61 for the positive @xmath47-projection , @xmath62 for the negative @xmath47-projection , @xmath63 for the positive @xmath48-projection , @xmath64 for the negative @xmath48-projection , @xmath65 for the positive @xmath49-projection and @xmath66 for the negative @xmath49-projection . in figs .
12 , we plot the marginal distributions of the @xmath0-spin particle @xmath67 the spin direction of which coincides with the positive direction of the @xmath47 axis and the spin projection onto a fixed axis in space is equal to @xmath68 , as a function of the angles @xmath29 and @xmath30 .
however , there exist the mixed states of the @xmath0-spin particle ( see , for example , @xcite ) .
these states are described only by the density matrix @xmath69 in eq .
( [ den!matrix!mixed ] ) , @xmath70 are determined by the relation @xmath71 , and @xmath72 , @xmath73 and @xmath74 are the mean values of the @xmath0-spin projections on the @xmath47 , @xmath48 and @xmath49 axes , respectively .
these parameters satisfy the condition @xmath75 .
a parameter @xmath76 is named `` purity state degree '' and depends on the mean values of the spin projection @xmath77 . as in the case of a pure state , using eq .
( [ diagon!elem ] ) for the matrix elements of the density matrix of a mixed state , we calculate the marginal distribution of the mixed state of the @xmath0-spin particle , @xmath78 and @xmath79 it is easy to check that the examples of the pure - state marginal distribution ( [ m!dis!x!pos])-([m!dis!z!op ] ) are obtained from eqs .
( [ m!dis!mixed1 ] ) , ( [ m!dis!mixed2 ] ) by the proper choice of the parameters @xmath72 , @xmath73 and @xmath74 . in this section
, one can also consider another one important example of the @xmath44-spin particle states , the direction of which coincides with the positive direction of the @xmath49 axis . in this particular case ,
the wave function has the form @xmath80 and the density matrix , consequently , takes the form @xmath81 the marginal distribution of the spin state can be calculated by the following equation ( see , for example , @xcite ) @xmath82 where @xmath17 is determined by eq .
( [ integration ] ) . as a result
, we obtain the marginal distribution @xmath83 , @xmath84 , i.e. , three probabilities for the @xmath44-spin projection onto the @xmath49 axes ( @xmath85 , @xmath86 , and @xmath87 ) , which also depend on the rotation angles of the reference frame @xmath30 , @xmath29 , and @xmath26 @xmath88 @xmath89 @xmath90 a dependence on the @xmath91 , @xmath30 angles drops out . substituting this marginal distribution into formula ( [ fig1 ] ) by means of which matrix elements of the density matrix @xmath92 can be reconstructed by the measurable marginal distribution @xmath93 and executing the calculations , we could verify the correction of this formula in the particular case of the unity spin . in the case
@xmath94 , this formula consists of twenty seven items .
let us consider all values of the wigner d - functions for the various values of their subscripts and superscripts . because the marginal distribution @xmath95 takes only positive values , and because some d - functions contain a complex - valued exponential coefficient , integration of the product of the marginal distribution and a d - function over the @xmath96 rotation variables yields zero .
taking this fact into account , eq .
( [ fig1 ] ) can be simplified , and we arrive at @xmath97 .
\end{array } \label{matrix!density2}\end{aligned}\ ] ] calculating all nonzero integrals and taking into account the expressions of the @xmath98-symbols from @xcite we can rewrite eq .
( [ matrix!density2 ] ) into the form @xmath99 .
\end{array}\end{aligned}\ ] ] for @xmath100 ( @xmath101 ) , @xmath98-symbols are equal to zero ( see , for example , @xcite ) . evaluating all the remaining @xmath98-symbols for @xmath102 , we obtain @xmath103 this means that we demonstrated how the initial density matrix ( [ m!d4 ] ) can be recovered , if we know the marginal distribution of the quantum state ( [ md1])-([md2 ] ) .
here we will consider the tomography scheme of rotational states of molecules and nuclei using , for simplicity , the model of a symmetric top .
it is a common knowledge ( see , for example , @xcite ) that the hamiltonian of a symmetric top has the form @xmath104 where @xmath105 , @xmath106 are the principal inertia momenta of a top ( two of the momenta of a symmetric top coincide with each other ) , and @xmath107 is the angular - momentum operator .
stationary states of a symmetric top are characterized by three quantum numbers : the orbital momentum @xmath32 and its projections onto the top axis ( @xmath108 ) and onto a fixed @xmath49 axis in space ( @xmath109 ) .
the symmetric top energy is independent of the last quantum number @xmath110 .
let us consider the stationary states of the symmetric top with given energy . to do this
, we take a subspace of the @xmath111 dimensions with the fixed @xmath32 in the state space .
the wave function of the stationary state of a symmetric top can be represented in the form @xmath112 where @xmath113 is the wigner d - function , and @xmath114 is the wave function in the reference frame , which comoving with the physical system ( the top ) . for a pure state ,
the density matrix of the symmetric top is expressed in terms of the wave functions as follows @xmath115 thus , under two consecutive rotations ( these rotations are determined by the euler angles @xmath116 and @xmath117 ) the hermitian density matrix @xmath118 is transformed as @xmath119 where unitary @xmath120 matrixes @xmath121 and @xmath122 represent these two rotations through the angles @xmath17 and @xmath123 .
since the density matrix under consideration depends on four discrete indices , it is necessary to do two consecutive rotations of the reference frame , which comoving with the top to obtain the diagonal elements of the transformed density matrix . taking into account
this fact , for the diagonal elements of the density matrix @xmath124 can be reduced to the form @xmath125 here @xmath126 is the above wigner d - function and @xmath127 . in eq .
( [ diagonal ] ) , we assume that summation is performed over the repeated indices @xmath128 , but there is no summation over the indices @xmath129 .
below , we will assume that the hermitian nonnegative density matrix @xmath130 describes not only the pure state of the top ( [ wave!function!top ] ) but also an arbitrary mixed state , i.e. @xmath131 .
let us discuss the problem of reconstructing all matrix elements @xmath130 , if the diagonal elements @xmath132 are known . to do this ,
let us multiply both sides of eq .
( [ diagonal ] ) by the functions @xmath133 and @xmath134 with arbitrary upper indices @xmath135 , and with different angular variables @xmath17 and @xmath123 . as a result
, we have @xmath136 integrating over the angular variables @xmath116 and @xmath137 , @xmath138 @xmath139 using the well - known expression of the integral of the three d - functions product @xcite and @xmath140 along with the orthogonality properties of the @xmath98-symbols , one can express @xmath130 in terms of the diagonal elements @xmath141 . as it was mentioned above , the diagonal elements of the density matrix for a quantum state take nonnegative values and their sum is equal to unity .
thus , we introduce the notation @xmath142 , where the function @xmath143 is the marginal distribution of a symmetric top , i.e. the probability of finding the projection @xmath144 of the angular momentum onto a fixed axis in space and the projection @xmath145 of the angular momentum onto the top axis , and also this probability parametrically depends on euler angles @xmath146 .
this function is normalized @xmath147 for all values of the euler angles .
suppose we know the positive , normalized marginal distribution @xmath148 .
thus , in view of eqs .
( [ times])-([flatly ] ) the matrix elements @xmath149 can be expressed in terms of the measurable marginal distribution @xmath143 with @xmath150 by using the relationship @xmath151 here @xmath152 .
one can conclude that given a measurable marginal distribution of the symmetric top , one can reconstruct the stationary - state density matrix of the symmetric top by means of eq .
( [ matrix!density ] ) .
s. mancini , v. i. manko and p. tombesi , _ phys .
a. _ , * 213 * , 1 , ( 1996 ) ; _ `` classical - like description of quantum dynamics by means of symplectic tomography '' _ , los alamos preprint quant - ph/9609026 . | a possibility of describing two - level atom states in terms of positive probability distributions ( analog to the symplectic tomography scheme ) is considered . as a result
the basis of the irreducible representation of a rotation group can be realized by a family of the probability distributions of the spin projection parametrized by points on the sphere .
furthermore the tomography of rotational states of molecules and nuclei which can be described by the model of a symmetric top is discussed .
* tomography of quantum spinor states + * * s. s. safonov + * _ moscow institute of physics and technology , _ + _ institutskii pr .
9 , dolgoprudny , moscow reg .
, 141700 , russia + and + p. n. lebedev physical institute , russian academy of science , + leninskii pr .
53 , moscow 117924 , russia _ recently @xcite the symplectic tomography scheme was suggested to obtain the wigner function by measuring the probability distribution function ( the marginal distribution ) for a shifted , rotated and squeezed quadrature , which depends on extra parameters .
the invertable map of the wigner function of a quantum state onto the positive marginal distribution for the continous observable ( coordinate ) was used to give the formulation of quantum dynamics as the classical statistical process @xcite . from this point of view
, the moyal approach @xcite to the quantum evolution as to a statistical process was improved in the sense that instead of the moyal quasidistribution function ( the wigner function ) was introduced the positive probability distribution of the measurable variables describing an arbitrary quantum state and its evolution .
a spin states reconstruction procedure similar to the symplectic tomography scheme is investigated in @xcite .
the authors of these papers showed that in the framework of the symplectic tomography approach it is possible to describe equivalently the spin state in terms of the classical distribution function of a discrete variable instead of the wave function or the density matrix .
taking into account the results of these papers in this article it will be investigated the tomography scheme for two - level atom states and rotational states of molecules and nuclei .
the goal of this work is to make a review of construction of the explicit formula for the invertable map connecting the quantum state of a two - level atom described by the @xmath0-spin density matrix @xmath1 with the probability distribution function @xmath2 of the @xmath0-spin projection on the quantization axis , where the angles @xmath3 , @xmath4 determine the coordinates of the vector normal to the surface of the sphere of the unity radius . on the other hand ,
quantum characteristics such as rotational energy levels of molecules and nuclei are connected with the tomographic description of the tops behavior . therefore in this work
it will be also discussed the classical - like description of the quantum states of a top in terms of the distribution functions .
for simplicity of presentation , we will consider the example of a symmetric top .
these results also are discussed in @xcite . |
the paramagnetic meissner effect ( pme ) observed in small superconducting samples simply expresses the fact that the energy of those samples decreases upon an increasing magnetic field . a reentrant ( h , t ) phase diagram is one of the signals that characterizes this effect .
this surprising effect has been seen both in conventional@xcite@xcite and in high tc superconductors @xcite .
the behavior of nanometer - scale superconducting grains@xcite , which shows pauli paramagnetism@xcite in addition to other striking properties of finite electron systems , is outside the scope of this work which relies on the ginzburg - landau equation .
metastability is one of the major ingredients that have been put forward to explain pme in conventional superconductors @xcite .
for example , in the ginzburg - landau approach of small size cylindrical samples at the upper critical field , the disbalance of the screening currents leads to a positive magnetic moment @xcite .
but , in this case if the equilibrium transitions are only allowed , the pme disapears @xcite , and many mechanisms causing non equilibrium flux configuration have been proposed @xcite .
however pme is known to occur for minimal energy configurations of some mesoscopic objects : the little - parks @xcite ring is the most simple example that shows an alternating sequence of diamagnetic and paramagnetic responses as the magnetic field is increased . on the theoretical side
this general behavior is known to be shared by all loop - like samples and moreover that pme appears the be stable against thermic fluctuations @xcite . in the framework of the london approximation of the vortex phase of small superconducting disks , we have obtained magnetization curves which showed a pme , in good agreeement with the experimental observation @xcite , and made the analogy with the little - parks ring to interpret such a behavior of stable configurations @xcite .
the aim of this paper is to give the reader a simple example of a mesoscopic ring - like superconducting film of which the paramagnetic response of the stable configurations can be tuned with the size of the central hole , and to show at which part of the ring ( inner or outer ) the nucleation of superconductivity is localized . in section 2
the upper critical field @xmath0 is calculated in the framework of the ginzburg - landau equation , and , if the size of the hole is large enough , the parts of the phase diagram where pme takes place are found . in section 3
we show the two regions of the ring with opposite supercurrents which might give rise to a global positive magnetic moment . in section 4
we give in more details our results for a ring with a small hole with @xmath1 for which there is only one paramagnetic domain and we also show the order parameter density profile that changes drastically when crossing the first order line .
our conclusions are drawn in the last section .
we shall follow quite closely the method of bezryadin , buzdin and pannetier @xcite in the case of a circular hole or of a disk@xcite , in order to obtain the phase diagram of a circular ring of external radius @xmath2 and width @xmath3 under a uniform magnetic field @xmath4 normal to the plane of the film sample . in the case of a thin film of thickness @xmath5 the effective screening length , @xmath6 , becomes very large and we may neglect the magnetic field energy in the ginzburg - landau ( gl ) free energy functional : @xmath7 where @xmath8 .
moreover due to the small size of the sample , we can take for the vector potential @xmath9 the potential of the uniform external applied field @xmath4 : @xmath10 using polar coordinates ( @xmath11 , @xmath12 ) , the equation for the normalized order parameter @xmath13 is : @xmath14 ^{2}\psi + \frac{r^{2}}{\xi^{2}}(\psi-\psi^{3})=0\ ] ] where @xmath15 @xmath16 , @xmath17 , @xmath18 , @xmath19 , and @xmath20 is the flux of the magnetic field across the plain disk ( with no hole ) . in order to obtain the upper critical field @xmath0 @xcite , using the linearized version of the gl equation , we search for solutions of definite angular momentum @xmath21 of the form : @xmath22 then , in the present gauge , the super - current is tangent to the two circular insulating boundaries of the ring , and we may take the boundary conditions of the gl equation as follows : @xmath23 after reduction to kummers s equation , the general solution can be expressed as the linear combination of two confluent hypergeometric functions @xmath24 and @xmath25 @xcite as follows : @xmath26 where @xmath27 , and where the two constants @xmath28 and @xmath29 have to be determined using the boundary conditions . in practice , we have first obtained the ratio @xmath30 in terms of @xmath31 , @xmath32 and @xmath21 using the first condition , while the second condition has been numerically solved and allows us to obtain @xmath33 versus @xmath31 for each @xmath21 .
then , for a given flux @xmath31 , the critical line @xmath34 is obtained by choosing the angular momentum value which minimizes @xmath32 @xcite
. the limiting cases of either a plain disk or a circular hole can be reached , using the same procedure , with @xmath35 in the first case and with @xmath36 in the second one .
the little - parks ring @xcite , can be recovered using a very small value of @xmath37 , as shown on figure 1 .
we see that paramagnetism starts just after the transition between angular momentum states @xmath38 and disapears for @xmath39 , until the next transition occurs . on figure 2
we give the @xmath0 line for a ring with @xmath40 .
the cusps , associated to jumps of the angular momentum , are clearly seen and three small paramagnetic domains are located above these transitions . for sake of comparison the full disk @xmath0 curve is shown on figure 3 , on which no such domains are seen ; such oscillations have been observed on micron - sized superconducting disks@xcite .
our results give @xmath41 as the value under which paramagnetic domains are observed .
in the state @xmath42 of angular momentum @xmath21 , the magnetic moment @xmath43 can be written as : @xmath44 in fact the local density of magnetic moment is related to the value of the supercurrent : @xmath45 of which the components are : @xmath46 for a given flux and angular momentum , and when the ring is crossed from the inner edge ( @xmath47 ) up to the outer one ( @xmath48 , three various situations can be encountered acording to the sign of @xmath49:@xmath50 * the radial supercurrent density is negative everywhere on the ring for @xmath51 .
this gives a uniform standard diamagnetic response .
* the radial supercurrent density is positive everywhere on the ring for @xmath52 and gives a uniform paramagnetic response * the current flows changes its sign at a point @xmath53 .
thus one can distinguish two different regions on the ring : the inner part which contibutes positively to the total magnetic moment and the outer part which gives a diamagnetic contribution , with currents in the two regions flowing in opposite directions .
let us comment figures 1 - 3 on this point .
the ideal little - parks ring of fig1 ( with @xmath54 ) shows a sequence of uniform responses : @xmath55 is diamagnetic , whereas the reentrant part of @xmath56 is paramagnetic , etc ... and the supercurrent changes its sign for all half - integer values of @xmath57 .
fig2 , which depicts the behavior of a thicker ring ( with @xmath58 ) , shows that apart the standard uniform diamagnetic response of the @xmath55 state , the @xmath56 state gives a uniform paramagnetic response for small fields @xmath59 .
the magnetization of all the other states results of the competition between the inner and the outer currents .
the case of the plain disk ( fig3 ) is interesting although no paragnetism is observed on the global magnetization : all the non - zero angular momentum states have a central paramagnetic response and a larger diamagnetic one located on the periphery of the disk .
in this section , we have choosen to treat more extensively the particular case of a small hole , @xmath60 , by solving the non - linear ginzburg - landau equation .
this allows us to enter the superconducting region , and compute the densities of both the order parameter and the magnetization .
in fact it can be shown that mixture of different angular momentum states can be neglected if only the lowest energy is searched for , so that we shall not consider such mixtures .
because the two boundary conditions restrict @xmath13 at two different points , namely @xmath47 and @xmath61 , we have solved the non - linear gl equation in two steps .
we have used standard numerical routines which allowed us to obtain @xmath62 given the initial values @xmath63 and @xmath64 , then @xmath63 is found by solving @xmath65 for @xmath63 . the @xmath0 line for the ring with @xmath60 is drawn on fig .
4 . only one small paramagnetic domain shows up for fluxes @xmath66 with @xmath56 . on the inset of fig 4 , which is a zoom of the paramagnetic region ,
we have drawn a fixed temperature path ( a - b ) with @xmath67 which crosses the first order transition line separating the @xmath55 and @xmath56 phases . a fixed flux path ( c - b ) with @xmath68 which lies entirely in the paramagnetic phase is also shown .
the field dependence of the magnetization versus flux along the path ( a - b ) is given on fig 5a , on which one sees the jump of the magnetic moment to a positive value when the first order line is crossed .
this behavior should be compared with the smooth one of the magnetization along path ( c - b ) , in the paramagnetic phase at fixed field and versus temperature , on fig 5b .
we now give on fig 6 the magnetization density at points a and b versus @xmath69 , i.e. when the ring is crossed from the inner circle to the outer one .
as already explained in the prevoius section , one clearly sees that the inner part is paramagnetic while the outer one is diamagnetic and that the balance of the two results in a small pme effect .
let us stress that , following this line of argument , and as said for disks in the previous section , one should observe for disks with very small holes a small domain with a positive local magnetic moment ( around the hole ) while there is no global pme effect .
we discuss now the order parameter distribution across the ring . on fig 7 , comparing these densities , we find that the order parameter is larger in the inner part than in the outer part , at point a i.e. in the diamagnetic region . on the contrary , in the paramagnetic region , at points
b , the order parameter is 30% larger in the vicinity of the outer edge than around the hole .
figure 8 shows how the order parameter density profile changes when crossing the first order transition at @xmath70 at fixed temperature @xmath67 .
we have plotted the values of @xmath71 on the two , inner and outer , boundaries of the ring for @xmath55 before the transition and for @xmath56 after .
it is worthwhile to notice that the regions of the ring where the order parameter is the largest , changes as the first order line is crossed , passing from the inner part before the transition to the outer one after .
we have checked on the example of @xmath40 that this result should be general . in this last case
it is observed that , sweeping the temperature range of @xmath56 from one transition point @xmath72 to the next one @xmath73 , the order parameter profile @xmath74 evolves from more populated on the inner edge to more populated on the outer one .
let us note that , even though the superconducting parameter is maximal in this last region , a paramagnetic effect arises because of the large inner part of the disk in which the magnetization density is positive ( see fig .
the paramagnetism of a mesoscopic object can obviously be due to any mechanism leading to non equilibrium configurations .
however , at equilibrium , the pme depends strongly on the geometry of the sample and presumably can be observed in many cases much more involded than the toy model presented here .
here we have shown that the giant vortex state of a pierced disk can exhibit pme in minimal energy configurations .
the size of the hole controls both the number and extension of the paramagnetic domains .
increase of the field at fixed temperaure to enter those domains causes the order parameter to jump from inner to outer localization .
the present approach does not explain the pme seen on the surface superconducting states of 1:4 elongated ellipses @xcite for which the pme in an asymmetric ring should be studied .
paramagnetic meissner effect in a multiply - connected array of josephson junctions has been reported @xcite for which the diamagnetic current flows on the exterior plaquettes wheras the paramagnetic current flows in the inside of the sample ; this is very similar to that we have described here .
a. buzdin is warmly thanked for his interest in this work and for numerous and helpfull discussions . 99 d.j .
thompson , m.s.m .
minhaj , l.e . wenger and j.t .
chen , phys .
lett . * 75 * , 529 ( 1995);p .
kostic , a.p .
paulikas , u. welp , v.r .
todt , c. gu , u. geiser , j.m .
williams , k.d .
carlson and r.a .
klemm , phys . rev . * b 53 * , 791 ( 1996);a .
terentiev , d.b .
watkins and l.e.d .
long , phys . rev . *
b 60 * , r761 ( 1999 ) .
w. braunisch , n. knauf , v. kataev , a. grutz , a. kock , b. roden , d. khomskii and d.wohlleben , phys .
lett . * 68 * , 1908 ( 1992 ) and w. braunisch , n. knauf , b. bauer , a. kock , a.becker , b. freitag , a. grutz , v. kataev , s. neuhausen , b. roden , d. khomskii and d.wohlleben , phys . rev .
* b 48 * , 4030 ( 1993);b .
schlipe , m. stindtmann , i. nikolic , and k. baberschke , phys . rev . * b 47 * , 8331 ( 1993);s .
riedling , g. brauchle , r. lucht , k. rohberg , and h. v. lohneysen , phys . rev . *
* b 4**9 , 13283 ( 1994);u .
onbasli , y.t .
wang , a. naziripour , r. tello , w. kielh , and a. m. hermann , phys . stat .
* b 194 * , 371 ( 1996);g.s .
okram , d.t .
adroja , b.d .
paladia , o. prakash , and p.a.j .
de groot , j. phys .
: cond . mat .
* 9 * , l525 ( 1997 ) .
a. e. koshelev and a. i. larkin , phys . rev . *
b 52 * ( 1995 ) 13559.v . v. moschalkov , x. g. qiu and v. bruyndoncx , phys . rev . *
b 55 * ( 1999 ) 11793.p .
s. deo , v. a. schweigert , f. m. peeters , and a. k. geim , phys .
* 79 * ( 1997 ) 4653.p .
s. deo , v. a. schweigert and f. m. peeters , phys . rev . *
b 59 * ( 1999 ) 6039 .
* figure 1*. the @xmath76phase diagram of a very thin ring with @xmath77 , versus normalized temperaure @xmath78 and flux @xmath79 in flux quantum units . the surface superconducting critical line @xmath80 has been drawn for the first three angular momentum values .
* figure 2 .
* the upper critical field @xmath80for a ring with @xmath58 .
the first order lines which separate domains of different angular momentum are drawn as thin full vertical lines , whereas vertical dashed lines mark the expected boundary of each paramagnetic region . the present study is valid to the right hand side of the @xmath81 line which has been shown for comparison ( for the bulk material ) .
* figure 4 . * the phase diagram of a disk with a small hole @xmath83 , in the vicinity of the upper critical fileld @xmath80 .
the inset plot shows a zoom of the region @xmath84 in which a paramagnetic effect is expected .
* figure 5a . * the magnetization at fixed normalized temperature @xmath85 versus flux along the path a - b ( see the inset graph of fig .
the magnetization jumps to small positive values after the transition . * figure 8 . * the order parameter density profile oscillation when crossing the first order transition at @xmath88 for a ring with @xmath83 .
the inner edge is more superconducting than the outer one before the transition , whereas the reverse situation occurs in the paramagnetic phase . | in the framework of the ginzburg - landau equation , the temperature dependence of the upper critical field of small ring - like superconductors is studied . at equilibrium small parts of the phase diagram show paramagnetism for width / radius ratios below 0.85 .
their number and extension increase with the size of the hole . in these regions ,
only the inner part of the ring shows a positive magnetic moment .
the order parameter density profile appears to change , when crossing a first order transition line , which separates different angular momentum values , and we clarify the relationship between the localization of superconductivity nucleation and paramagnetism of those samples . |
the mrfm consists of an nanomagnetic tip integrated on top of a microwire rf source , an ultrasensitive si cantilever , and a fiber - optic interferometer to measure its displacement @xcite .
the entire apparatus operates in vacuum better than @xmath83 mbar , at temperatures down to @xmath73 k , and in an applied longitudinal field up to @xmath84 t. the nw of interest is attached to the end of the cantilever and is positioned within 100 nm of the nanomagnetic tip , as shown schematically in fig .
[ fig1 ] . here
the field produced by the tip , @xmath85 , results in a total static magnetic field @xmath86 , while the microwire produces a transverse rf field @xmath87 .
arp pulses invert the nuclear spin polarization because the polarization follows ( or is `` spin - locked '' to ) the time - dependent effective field @xmath88 in a frame rotating with the rf field , where @xmath89 is the instantaneous frequency of the arp pulses , @xmath90 their amplitude , and the unit vectors are defined in the rotating frame .
the volume of inverted spins , known as the `` resonant slice '' , is determined by the spatial dependence of @xmath85 and the parameters of the pulses .
this force variance is in turn related to the number of nuclear spins in this volume by the approximation , @xmath91 , where @xmath92 is the direction of cantilever oscillation @xcite .
this relation holds as long as the volume of spins is small enough that the gradient @xmath93 is nearly constant throughout ; otherwise it provides a lower bound @xmath94 on the number of spins . from measurements of the magnetic field gradient in the vicinity of the tip
, we estimate that @xmath95 t / m at the position of the detection volume ( see supplementary material ) . in order to set an upper bound for the number of spins @xmath96
, we make a model of the magnetic field profile produced by the nanomagnetic tip and numerically integrate the spatially dependent gradient over the sample volume contained in the `` resonant slice '' ( see supplementary material ) .
ultrasensitive cantilevers are made from undoped single - crystal si and measure 120 @xmath97 m in length , 4 @xmath97 m in width , and 0.1 @xmath97 m in thickness . in vacuum and at the operating temperatures ,
the nw - loaded cantilevers have resonant frequencies @xmath98 and @xmath99 khz , intrinsic quality factors @xmath100 and @xmath101 , and spring constants @xmath102 and @xmath103 @xmath97n / m for the inp and gap nw experiments respectively .
the inp nw is 8 @xmath97m - long ; its diameter shrinks from 200 nm to 60 nm along its length ; and it is tipped by a 60 nm diameter au catalyst particle , left over from the growth process .
the gap nw is 10 @xmath97m - long ; its diameter is 1.0 @xmath97 m ; and it has a 1.5-@xmath97m - long tapered tip which reaches 90 nm in diameter at the au droplet . in this case
we remove the au droplet by cutting of the end of the nw with a focused ion beam , resulting in a 300-nm - diameter gap tip . finally , we sputter a 5 nm layer of pt onto the nw in order to shield electrostatic interactions @xcite .
each nw is affixed to the end of the cantilever with less than 100 fl of epoxy ( gatan g1 ) . in the attachment process
, we employ an optical microscope equipped with a long working distance and a pair of micromanipulators . during the measurement , the cantilever
is actively damped in order to give it a fast response time of @xmath104 ms .
up to 50 nw of laser light at 1550 nm are incident on the cantilever as part of the fiber - optic interferometer .
the nanomagnetic tips are truncated cones of dy fabricated by optical lithography @xcite . for the inp ( gap )
nw experiment the tip measures 225 ( 280 ) nm in height , 270 ( 250 ) nm in upper diameter , and 380 ( 500 ) nm in lower diameter . the rf source , on which the dy tip sits , is a 2-@xmath97m - long , 1-@xmath97m - wide , and 200-nm - thick au microwire . by positioning each nw within 100 nm of the combined structure
, the detection volume can be exposed to spatial magnetic field gradients exceeding @xmath105 t / m and rf @xmath90 fields larger than 20 mt without significant changes in the experimental operating temperature .
we use hyperbolic secant arp pulses with @xmath106 and a modulation amplitude set to 500 khz peak - to - peak for @xmath18p and 1 mhz for @xmath11h @xcite .
these parameters , combined with the geometry of the sample and the profile of @xmath93 , determine the size of the detection volume @xcite .
the authors thank c. l. degen , c. klffel , t. poggio , and r. j. warburton for illuminating discussions ; hari s. solanki for experimental assistance ; and s. keerthana for assistance with a figure .
we acknowledge support from the canton aargau , the swiss national science foundation ( snf , grant no .
200020 - 140478 ) , the swiss nanoscience institute , and the national center of competence in research for quantum science and technology .
p.p . and m.p conceived and planned the experiments in collaboration with f.x .
p.p . carried out the experiments .
p.p . and m.p .
analyzed the data and wrote the manuscript .
f. x. prepared the samples and devices .
the nanowires were grown by h.i.t.h .
, s.a . and e.p.a.m.b .
all authors discussed the results and contributed to the manuscript .
@xmath18p spins in an inp nw .
* ( a ) in - phase @xmath22 ( black ) and quadrature @xmath21 ( light gray ) force on the cantilever demodulated at the cantilever frequency .
@xmath73 k and @xmath84 t. the lock - in time constant is @xmath107 s , in order to match the correlation time @xmath108 s of the spin fluctuations and to reject the cantilever s thermal fluctuations with a correlation time @xmath104 ms .
the thermal noise @xmath21 sets the limit for the mrfm detection sensitivity at a polarization equivalent to @xmath109 @xmath18p nuclear spins rms .
( b ) histograms of @xmath22 and @xmath21 recorded for 1 hour .
gaussian fits to these histograms give @xmath110 an and @xmath111 an . ]
-inversion pulses .
( b ) @xmath22 recorded with a rectification threshold of -5 an ( top ) and with an absolute value threshold of 12.5 an ( bottom ) .
green pulses show the times at which @xmath36-inversions are applied .
@xmath112 s in both cases .
the sample is an ensemble of @xmath113 @xmath18p spins in an inp nw at @xmath73 k and @xmath84 t. ( c ) histograms of @xmath22 recorded over 1 hour corresponding to the natural ( black ) , rectified ( red ) , and narrowed ( blue ) cases .
the mean polarization of the rectified distribution is @xmath114 an compared with @xmath115 an of the natural distribution .
the standard deviation of the narrowed distribution is @xmath116 an , compared with @xmath117 an of the natural distribution . ]
averaged over 100 capture - store - readout sequences showing @xmath118 s for a @xmath18p spin ensemble with @xmath119 @xmath18 in a gap nw .
the mean stored fluctuation @xmath59 is shown as a filled diamond .
@xmath73 k and @xmath84 t. the signal s decay after the capture is due to the lock - in time constant @xmath120 s. the read - out fluctuations decay over a time @xmath121 s. bottom : @xmath122 averaged over the same 100 sequences showing the narrowed variance of the retrieved fluctuation @xmath123 ( filled diamond ) as it equilibrates back to its rotating frame value @xmath76 ( dotted line ) . ] for an ensemble of @xmath124 @xmath18p spins in an inp nw
. fits to @xmath22 during the readout take into account the lock - in time constant @xmath112 s ( solid lines ) and allow us to recover the value of the retrieved fluctuation @xmath125 ( filled circles ) and its exponential decay with @xmath126 s without the effect of @xmath58 ( solid lines ) .
the filled diamond indicates the mean stored fluctuation @xmath59 .
( b ) the same measurement done for @xmath127 @xmath11h spins on a gap nw with @xmath128 ms where @xmath17 is found to be @xmath129 ms .
again @xmath130 are displayed as filled circles and @xmath131 as a filled diamond . in both cases
@xmath73 k and @xmath84 t. ] bloch , f. nuclear induction .
_ * 70 * , 460 - 474 ( 1946 ) ] .
sleator , t. , hahn , e. l. , hilbert , c. & clarke , j. nuclear - spin noise . http://dx.doi.org/10.1103/physrevlett.55.1742[_phys
_ * 55 * , 1742 - 1745 ( 1985 ) ] ; sleator , t. & hahn , e. l. nuclear - spin noise and spontaneous emission .
b _ * 36 * , 1969 - 1980 ( 1987 ) ] .
degen , c. l. , poggio , m. , mamin , h. j. , rettner , c. t. & rugar , d. nanoscale magnetic resonance imaging .
usa _ * 106 * , 1313 - 1317 ( 2009 ) ] .
mamin , h. j. , kim , m. , sherwood , m. h. , rettner , c. t. , ohno , k. , awschalom , d. d. & rugar , d. nanoscale nuclear magnetic resonance with a nitrogen - vacancy spin sensor .
http://dx.doi.org/10.1126/science.1231540[_science_ * 339 * , 557 - 560 ( 2013 ) ] .
staudacher , t. , shi , f. , pezzagna , s. , meiger , j. , du , j. , meriles , c. a. , reinhard , f. & wrachtrup , j. nuclear magnetic resonance spectroscopy on a ( 5-nanometer)@xmath132 sample volume .
* 339 * , 561 - 563 ( 2013 ) ] .
nichol , j. m. , naibert , t. r. , hemesath , e. r. , lauhon , l. j. & budakian , r. nanoscale fourier - transform mri of spin noise . http://arxiv.org/abs/1302.2977[_arxiv_:1302.2977 ( 2013 ) ] .
mccoy , m. a. & ernst , r. r. nuclear spin noise at room temperature .
lett . _ * 159 * , 587 - 593 ( 1989 ) ] .
guron , m. & leroy , j. l. nmr of water protons .
the detection of their nuclear - spin noise , and a simple determination of absolute probe sensitivity based on radiation damping .
reson . _ * 85 * , 209 - 215 ( 1989 ) ] .
mamin , h. j. , budakian , r. , chui , b. w. & rugar , d. detection and manipulation of statistical polarization in small spin ensembles .
lett . _ * 91 * , 207604 ( 2003 ) ] ; mamin , h. j. , budakian , r. , chui , b. w. & rugar , d. magnetic resonance force microscopy of nuclear spins : detection and manipulation of statistical polarization .
b _ * 72 * , 024413 ( 2005 ) ] .
mller , n. & jerschow , a. nuclear spin noise imaging .
natl . acad .
usa _ * 103 * , 6790 - 6792 ( 2006 ) ] .
merkulov , i. a. , efros , ai .
l. & rosen , m. electron spin relaxation by nuclei in semiconductor quantum dots .
b _ * 65 * , 205309 ( 2002 ) ] .
khaetskii , a. v. , loss d. & glazman , l. electron spin decoherence in quantum dots due to interaction with nuclei .
* 88 * , 186802 ( 2002 ) ] .
childress , l. , gurudev dutt , m. v. , taylor , j. m. , zibrov , a. s. , jelezko , f. , wrachtrup , j. , hemmer , p. r. & lukin , m. d. coherent dynamics of coupled electron and nuclear spin qubits in diamond .
http://dx.doi.org/10.1126/science.1131871[_science_ * 314 * , 281 - 285 ( 2006 ) ] .
kuhlmann , a. v. , houel , j. , ludwig , a. , greuter , l. , reuter , d. , wieck , a. d. , poggio , m. & warburton , r. j. charge noise and spin noise in a semiconductor quantum device .
http://arxiv.org/abs/1301.6381[_arxiv_:1301.6381 ( 2013 ) ] .
coish w. a. & loss d. hyperfine interaction in a quantum dot : non - markovian electron spin dynamics .
b _ * 70 * , 195340 ( 2004 ) ] .
reilly , d. j. , taylor , j. m. , petta , j. r. , marcus , c. m. , hanson , m. p. & gossard , a. c. suppressing spin qubit dephasing by nuclear state preparation .
http://dx.doi.org/10.1126/science.1159221[_science_ * 321 * , 817 - 821 ( 2008 ) ] .
latta , c. , hgele , a. , zhao , y. , vamivakas , a. n. , maletinsky , p. , kroner , m. , dreiser , j. , carusotto , i. , badolato , a. , schuh , d. , wegscheider , w. , atature , m. & imamoglu , a. confluence of resonant laser excitation and bidirectional quantum - dot nuclear - spin polarization .
http://dx.doi.org/10.1038/nphys1363[_nature phys . _
* 5 * , 758 - 763 ( 2009 ) ] .
vink , i. , nowack , k. c. , koppens , f. h. l. , danon , j. , nazarov , y. v. & vandersypen , l. m. k. locking electron spins into magnetic resonance by electron - nuclear feedback .
http://dx.doi.org/10.1038/nphys1366[_nature phys . _
* 5 * , 764 - 768 ( 2009 ) ] .
bluhm , h. , foletti , s. , mahalu , d. , umansky , v. & yacoby , a. enhancing the coherence of a spin qubit by operating it as a feedback loop that controls its nuclear spin bath .
lett . _ * 105 * , 216803 ( 2010 ) ] .
< togan , e. , chu , y. , imamoglu , a. & lukin , m. d. laser cooling and real - time measurement of the nuclear spin environment of a solid - state qubit .
http://dx.doi.org/10.1038/nature10528[_nature_ * 478 * , 497 - 501 ( 2011 ) ] .
xue , f. , weber , d. p. , peddibhotla , p. & poggio , m. measurement of statistical nuclear spin polarization in a nanoscale gaas sample .
b _ * 84 * , 205328 ( 2011 ) ] .
budakian , r. , mamin , h. j. , chui , b. w. & rugar , d. creating order from random fluctuations in small spin ensembles .
http://dx.doi.org/10.1126/science.1106718[_science_ * 307 * ,
408 - 411 ( 2005 ) ] .
assali , s. , zardo , i. , plissard , s. , kriegner , d. , verheijen , m. a. , bauer , g. , meijerink , a. , belabbes , a. , bechstedt , f. , haverkort , j. e. m. & bakkers , e. p. a. m. http://dx.doi.org/10.1021/nl304723c[_nano lett .
_ asap , doi : 10.1021/nl304723c ] .
tanns , a. & garwood , m. improved performance of frequency - swept pulses using offset - independent adiabaticity .
reson . , ser a _ * 120 * , 133 - 137 ( 1996 ) ] ; tomka , i. t. , van beek , j. d. , joss , r. & meier , b. h. spatio - chemical characterization of a polymer blend by magnetic resonance force microscopy .
_ * 15 * , 3438 ( 2013 ) ] .
degen , c. l. , poggio , m. , mamin , h. j. & rugar , d. nuclear spin relaxation induced by a mechanical resonator .
* 100 * , 137601 ( 2008 ) ] .
rice , s. o. mathematical analysis of random noise .
_ at&t tech .
j. _ * 23 * , 282 - 332 ( 1944 ) ; _ at&t tech . j. _ * 24 * , 46 - 156 ( 1944 ) .
degen , c. l. , poggio , m. , mamin , h. j. & rugar , d. role of spin noise in the detection of nanoscale ensembles of nuclear spins .
lett . _ * 99 * , 250601 ( 2007 ) ] .
poggio , m. , degen , c. l. , rettner , c. t. , mamin , h. j. & rugar , d. nuclear magnetic resonance force microscopy with a microwire rf source .
* 90 * , 263111 ( 2007 ) ] .
stipe , b. c. , mamin , h. j. , stowe , t. d. , kenny , t. w. & rugar , d. noncontact friction and force fluctuations between closely spaced bodies .
lett . _ * 87 * , 096801 ( 2001 ) ] .
mamin , h. j. , rettner , c. t. , sherwood , m. h. , gao , l. & rugar , d. high field - gradient dysprosium tips for magnetic resonance force microscopy .
lett . _ * 100 * , 013102 ( 2012 ) ] . | soon after the first measurements of nuclear magnetic resonance ( nmr ) in a condensed matter system , bloch @xcite predicted the presence of statistical fluctuations proportional to @xmath0 in the polarization of an ensemble of @xmath1 spins .
first observed by sleator et al .
@xcite , so - called `` spin noise '' has recently emerged as a critical ingredient in nanometer - scale magnetic resonance imaging ( nanomri ) @xcite .
this prominence is a direct result of mri resolution improving to better than @xmath2 , a size - scale in which statistical spin fluctuations begin to dominate the polarization dynamics .
we demonstrate a technique that creates spin order in nanometer - scale ensembles of nuclear spins by harnessing these fluctuations to produce polarizations both larger and narrower than the natural thermal distribution .
we focus on ensembles containing @xmath3 phosphorus and hydrogen spins associated with single inp and gap nanowires ( nws ) and their hydrogen - containing adsorbate layers .
we monitor , control , and capture fluctuations in the ensemble s spin polarization in real - time and store them for extended periods .
this selective capture of large polarization fluctuations may provide a route for enhancing the weak magnetic signals produced by nanometer - scale volumes of nuclear spins .
the scheme may also prove useful for initializing the nuclear hyperfine field of electron spin qubits in the solid - state .
spin noise is a phenomenon present in all spin ensembles that begins to exceed the mean thermal polarization as the size of the ensemble shrinks .
these fluctuations have random amplitude and phase and have been observed in a wide variety of nuclear spin systems including in liquids by conventional nmr @xcite and in the solid - state using a superconducting interference device ( squid)@xcite , by force - detected magnetic resonance @xcite , or by nitrogen - vacancy ( nv ) magnetometry @xcite .
in addition to finding applications in nanomri , spin noise has been used in mri specially adapted for the investigation of extremely delicate specimens , in which external radio frequency ( rf ) irradiation is not desired @xcite . through the hyperfine interaction ,
nuclear spin noise also sets limits on the coherence of electronic qubits in the solid state @xcite .
various efforts to mitigate these nuclear field fluctuations by hyperfine - mediated nuclear spin preparation have been developed , both in quantum dots ( qds ) @xcite and in nv centers in diamond @xcite .
we report on a manipulation and initialization technique which applies to arbitrary nanometer - scale samples , i.e. it does not require specialized structures providing a controllable electronic spin and a strong hyperfine interaction .
rather , our technique requires a detector sensitive enough to resolve nuclear spin noise and the ability to apply rf electro - magnetic pulses .
conventional nmr and mri techniques rely on manipulating the mean thermal polarization to produce signals .
statistical nuclear polarization fluctuations exceed the mean thermal polarization below a critical number of spins @xmath4 , where @xmath5 is the spin quantum number , @xmath6 is the boltzmann constant , @xmath7 is the temperature , @xmath8 is planck s constant , @xmath9 is the gyromagnetic ratio , and @xmath10 is the magnetic field @xcite . for @xmath11h spins at @xmath12
k in an applied magnetic field @xmath13 t , @xmath14 , which in most organic samples corresponds to a critical volume @xmath15 . under the ambient conditions of recent nmr experiments with nv spin sensors , statistical polarization begins to dominate for even larger samples with @xmath16 @xcite . in order to measure and initialize such ensembles ,
it is therefore worthwhile to consider techniques designed to both detect and control nuclear polarization fluctuations .
although nuclear magnetic signals from such small volumes are weak , if a detector is able to resolve spin noise in real - time , i.e. faster than the spin correlation time @xmath17 , large fluctuations can be captured .
once captured , these fluctuations can then be used to initialize the polarization of nanometer - scale spin ensembles with a fixed sign and magnitude .
such initialization schemes could provide the basis for enhancing signals from small samples and for realizing advanced pulse protocols which can be borrowed directly from conventional nmr . in 2005 ,
budakian et al .
first demonstrated a technique for the capture and storage of electron spin fluctuations @xcite .
inspired by this work , we have adapted the protocol to nanometer - scale ensembles of nuclear spins . given
the relevance of these ensembles to nanomri and to the decoherence of solid - state qubits , this implementation could find wide applicability .
furthermore , the longer spin relaxation time of nuclear spins compared to electron spins sometimes as long as several hours at cryogenic temperatures allows captured spin order to be stored for far longer times .
we study two separate samples : an inp and a gap nw , both grown with the vapor - liquid - solid method in a metal - organic vapor - phase epitaxy reactor using gold droplets as catalyst @xcite . using a magnetic resonance force microscope ( mrfm ) , we measure the polarization of nanometer - scale ensembles of @xmath18p nuclei within each nw and of @xmath11h nuclei contained in the hydrocarbon adsorbate layer on the surface .
adiabatic rapid passage ( arp ) pulses of the transverse field @xcite are used to cyclically invert the polarization of nanometer - scale volumes of a particular nuclear isotope . in a magnetic field gradient
, these periodic inversions generate an alternating force that drives the mechanical resonance of an ultrasensitive cantilever ; the resulting oscillations , which are proportional to this force , are detected by a fiber - optic interferometer .
we feed the cantilever force signal to a lock - in amplifier referenced to the periodic spin inversions and monitor its in - phase ( @xmath19 ) and quadrature ( @xmath20 ) amplitudes .
the limiting source of noise in the measurement is the thermal noise of the cantilever , which has a random phase and on average contributes equally to @xmath19 and @xmath20 .
on the other hand , force fluctuations due to the nuclear spin polarization are in phase with the inversion pulses and contribute only to @xmath19 .
thus , as shown in fig .
[ fig2]a , we can monitor @xmath21 , the thermal force , and @xmath22 , the thermal force plus the force due to the nuclear spin inversions .
@xmath22 is dominated by the large fluctuations and the long @xmath17 of the spin noise , while the thermal noise in @xmath21 has a smaller amplitude and a shorter correlation time set by the damped cantilever force sensor ( see methods ) .
@xmath17 is limited by the magneto - mechanical noise originating from the thermal motion of the cantilever in a magnetic field gradient and by the arp pulse parameters @xcite . since the contribution of the spin signal to @xmath22 is large enough , we can follow in real - time the instantaneous nuclear spin imbalance in the rotating frame .
fig .
[ fig2]b shows the variances @xmath23 and @xmath24 , which give the variance due only to the thermal noise @xmath25 and the variance due only to the spin noise , @xmath26 .
@xmath27 is in turn related to the number of nuclear spins in the detection volume ( see methods ) .
the typical size of the spin ensembles we measure is between @xmath28 and @xmath29 for @xmath11h and between @xmath30 and @xmath31 for @xmath18p . given the density of typical adsorbate layers and inp and gap , the detection volumes discussed here are between @xmath32 and @xmath33 for @xmath11h and between @xmath34 and @xmath35 for @xmath18p .
the real - time measurement of spin noise allows us to react to and control the fluctuations in polarization @xcite . as a demonstration
, we conditionally apply rf @xmath36-inversion pulses to both rectify and narrow the naturally occurring polarization distribution .
this feedback is realized via a field - programmable gate array ( fpga ) in conjunction with an arbitrary waveform generator ( awg ) . as shown in fig .
[ fig3]a , when @xmath22 exceeds a predetermined threshold , the protocol applies a @xmath36-inversion , by inserting an arp pulse with a duration equivalent to a full cantilever cycle rather than the usual half - cycle . as a result , the spin ensemble s periodic inversion at the cantilever frequency undergoes a @xmath37 phase shift . in this way
, we can either rectify or narrow @xmath22 by setting appropriate thresholds as shown in fig .
[ fig3]b .
since only fluctuations due to nuclear spins are affected by the @xmath36-inversions , the effectiveness of the protocol depends on the fraction of @xmath22 arising from spin compared to thermal fluctuations , i.e. the larger the power signal - to - noise ratio ( snr ) @xmath38 is , the more effective the control of @xmath22 will be . as shown in fig .
[ fig3]c , this process can produce both hyper - polarized and narrowed nuclear spin distributions in the rotating frame .
in addition to controlling an ensemble s natural spin fluctuations , we can also capture especially large fluctuations @xcite . as shown in fig .
[ fig4 ] , using the fpga and awg , we continuously measure the rotating - frame spin fluctuations by monitoring @xmath22 .
once @xmath22 reaches a predetermined threshold @xmath39 , the arp pulses are tuned out of resonance with the nuclear spin ensemble , leaving the instantaneous spin polarization pointing along @xmath40 and transferring the transient spin order to the laboratory frame .
since the spin ensemble does not respond to the off - resonant arp pulses , the cantilever is no longer driven by spin forces . in this way
, hyper - polarized states of the nuclear spin ensemble can be captured and stored in the laboratory frame for as long as @xmath41 .
crucially , the time - scale required to capture the spin order here given by the inverse of the cantilever frequency @xmath42 must be much shorter than the rotating - frame correlation time @xmath17 . to confirm that the nuclear polarization has in fact been stored
, we can reapply the resonant arp pulses after some storage time @xmath43 to readout the polarization , once again cyclically inverting the polarization .
these spin inversions again drive the cantilever motion , and the amplitude of @xmath22 reflects the size of the retrieved fluctuation . in an idealized case , in which the spin component of the captured fluctuation @xmath39 is fully projected onto @xmath40 , the stored fluctuation @xmath44 will be normally distributed with a mean @xmath45 and a variance @xmath46 ( see supplementary material ) .
the fact that @xmath47 reflects the finite snr of the measurement , in this case limited by the cantilever s thermal noise .
note also that the distribution of the stored polarization is narrowed , i.e. it has a reduced variance , compared to the variance of @xmath22 under normal evolution ( @xmath48 ) . in the limit of large snr ( @xmath49 ) , @xmath50 and @xmath51 .
if during @xmath43 this stored polarization undergoes negligible relaxation in the laboratory frame , the corresponding retrieved fluctuation @xmath52 has a mean @xmath53 and a variance @xmath54 .
note that @xmath55 due to the finite snr of the retrieval measurement . in order to compare our experiments to this idealized case , we measure @xmath56 and @xmath57 for nanometer - scale @xmath18p and @xmath11h spin ensembles . as shown in fig .
[ fig5 ] , @xmath56 and @xmath17 are extracted from bi - exponential fits to @xmath22 during the readout sequence using our knowledge of the lock - in time constant @xmath58 .
deviations of @xmath56 from @xmath59 could be caused by the spin - lattice relaxation of the polarization in the laboratory frame
set by @xmath41 or by the incomplete projection of the polarization onto @xmath40 . for both @xmath18p and @xmath11h in fig .
[ fig5]a and b , these deviations are negligible within our error for two values of @xmath43
. limitations of the experimental hardware prevented measurements for larger @xmath43 , although our data show that @xmath60 s for @xmath18p in inp and @xmath61 s for @xmath11h on gap at @xmath62 k. fig .
[ fig4]b shows the reduced variance of the retrieved polarization @xmath57 , eventually approaching @xmath23 after a time on the order of @xmath17 in the rotating frame . note that @xmath58 is much shorter than the time over which @xmath57 evolves , excluding the lock - in as a source of the behavior .
this result demonstrates that the distribution of the captured fluctuations is indeed narrowed relative to the natural distribution of the nuclear spin fluctuations .
the ability to prepare such distributions may find application in quantum information processing with solid - state electron spins , which is often limited by the random nuclear field distribution in the host material @xcite .
the nuclear spin ensemble in such a system could be initialized before each measurement by a scheme based on the capture of random polarization fluctuations , thus enhancing the electron - spin dephasing time .
the degree of narrowing demonstrated in fig .
[ fig4]b represents a factor of @xmath63 compared to the natural distribution and is limited by the snr of the measurement .
the size of the captured spin fluctuation is , in principle , only limited by the amount of time one is willing to wait during the capture step . for the normally distributed random variable @xmath22 , the average amount of wait time required to capture a fluctuation @xmath39 is given by @xmath64 , where @xmath65 is the average number of times @xmath22 crosses zero per second ( see supplementary material ) @xcite .
for example , given that @xmath66 hz is a typical value in our experiments , fluctuations of @xmath67 are expected after just @xmath68 min , or alternatively for @xmath69 hr , a fluctuation of @xmath70 can be expected . for a @xmath18p spin ensemble with @xmath71 at @xmath72 t and @xmath73 k ,
the standard deviation of the statistical polarization fluctuations is given by @xmath74 and its mean thermal polarization is @xmath75 @xcite .
therefore , in the limit of large snr where @xmath76 is dominated by spin fluctuations , we can expect to capture polarizations of @xmath77 in 15 min and @xmath78 in 1 hr .
the polarization captured in a given @xmath79 could be increased even further by reducing @xmath17 and therefore increasing @xmath65 , e.g. through the periodic randomization of the spin ensemble using bursts of @xmath80-pulses @xcite . in principle , the nuclear spin decoherence time @xmath81 sets the lower limit for @xmath17 . as the size of the spin ensemble shrinks , the size of the achievable polarization increases as @xmath0 , making such a protocol increasingly relevant as detection volumes continue to shrink .
given that conventional pulse protocols based on thermal polarization require an initialization step taking at least @xmath41 , waiting for an extended time to capture a large spin fluctuation may be attractive especially when the magnitude of the captured fluctuation greatly exceeds the possible thermal polarization .
note that the crucial step in the capture protocol is that the monitoring and capture occurs in the rotating frame , where the time between statistically independent spin configurations set by @xmath17 is much faster than the equivalent time @xmath41 in the laboratory frame .
this reduced rotating - frame correlation time , allows the system to quickly explore its spin configuration space . on the other hand ,
once a large fluctuation is captured and transferred to the laboratory frame , the long @xmath41 effectively freezes the ensemble in this rare configuration .
we therefore show the long - term storage of large polarization fluctuations arising from spin noise in nanometer - scale nuclear spin ensembles with @xmath82 .
storage times as long as 20 s for @xmath18p are demonstrated , limited only by the measurement hardware . the ultimate limit to these times
is set only by @xmath41 in the laboratory frame , which at low temperatures is typically extremely long , exceeding hours for some nuclear spins .
while these results were obtained with a low - temperature mrfm , the capture and storage of spin fluctuations should be generally applicable to any technique capable of detecting and addressing nanometer - scale volumes of nuclear spins in real - time .
the ability to initialize the polarization of a small nuclear spin ensemble is important for the development of future nanometer - scale nmr techniques or possibly for the implementation of solid - state qubits and nuclear spin memory devices . given that the mean thermal polarization is typically negligible in nanometer - scale samples , other methods to create nuclear spin order must be considered . when polarization can not be created via effects such as overhauser - mediated dynamic nuclear polarization or electron - nuclear double resonance ,
the selective capture of large statistical fluctuations provides a viable alternative .
the ensembles polarized in this work , nanometer - scale volumes of @xmath11h on an adsorbate layer and @xmath18p in a single semiconducting nw , demonstrate the types of samples which could benefit from this technique .
one could imagine , for instance , such nuclear polarization capture processes enhancing the weak mri signals of a nanometer - scale @xmath11h - containing biological sample on a surface or of a semiconducting nanostructure . |
many results in quantum information theory require the generation of specific quantum states , such as epr pairs , or the implementation of specific quantum measurements , such as a von neumann measurement in a fourier transformed basis . some states and measurements can be efficiently implemented using standard quantum computational primitives such as preparing a qubit in the state @xmath0 and applying a sequence of quantum gates ( from a finite set ) .
epr pairs can be prepared from the state @xmath1 using a hadamard gate and a controlled - not gate .
a von neumann measurement in the fourier basis can be efficiently realized by applying an inverse quantum fourier transform and performing a von neumann measurement in the standard computational basis , i.e. @xmath2 .
however many states and basis changes can not be efficiently realized .
this paper focusses on the generation of quantum states .
for example , in @xcite , their improved frequency standard experiment requires the preparation of specific symmetric states on @xmath3 qubits , where @xmath3 is a parameter ( number of ions ) .
the algorithm we describe here will efficiently prepare the required symmetric state .
this short paper will focus on the algorithm for generating the state , and will ignore issues related to errors and decoherence ( for which the theory of fault - tolerant error - correction , or other stabilization methods , will apply ) .
we do not have space to elaborate on the details of precision , but simple calculations that require @xmath4 extra space and @xmath5 elementary operations allow us to generate any state with fidelity at least @xmath6 .
suppose we want to generate the state @xmath7 . in practice
it suffices to generate a state @xmath8 satisfying @xmath9 for a given small real number @xmath10 . in this case , it suffices to approximate each @xmath11 and @xmath12 to accuracy @xmath13 ( i.e. @xmath14 bits of accuracy ) .
here we have factored out the phase in each term , and so the @xmath12 are all non - negative real values .
note that if we can prepare the state @xmath15 , then we can approximate @xmath16 arbitrarily well by introducing appropriate phase factors using methods discussed in @xcite .
we will therefore focus on a method for generating states @xmath16 with non - negative real amplitudes .
in order to create the @xmath3-qubit state @xmath16 , we will implement in sequence @xmath3 controlled rotations , with the @xmath17th rotation controlled by the state of the previous @xmath18 qubits for @xmath19 .
we will first define these controlled rotations , and then in the next section we will describe how we would implement them . will extend the definition of @xmath12 to @xmath20 for @xmath21 .
suppose we had a copy of @xmath16 , and we measured the leftmost @xmath22 qubits in the computational basis .
let @xmath12 be the non - negative real number so that @xmath23 equals the probability the measurement result is @xmath24 .
then @xmath25 gives the conditional probability that the @xmath17th qubit is @xmath0 , conditioned on the state of the first state of the @xmath18 qubits being @xmath26 .
define a controlled rotation @xmath27 by : @xmath28 as shown in figure [ genpsi.fig ] , the algorithm for generating the @xmath3-qubit state @xmath29 is a sequence of @xmath3 such controlled rotations .
it is easy to show by induction that after the first @xmath22 controlled rotations are applied we have produced the state @xmath30 and therefore after all @xmath3 controlled rotations we have @xmath31 , width=240,height=144 ]
in this section we show how to implement the controlled-@xmath32 with arbitrary precision .
first assume we have a quantum register @xmath33 which encodes some `` classical '' description of the state @xmath34 .
the state @xmath33 must contain enough information to allow the probabilities @xmath35 ( or a related quantity , such as the @xmath36 we define below ) to be computed .
we also use an ancilla register of @xmath4 qubits initialized to the state @xmath0 .
then we define operators @xmath37 for each @xmath38 as follows : @xmath39 where @xmath40 satisfies @xmath41 a simple application of the techniques in @xcite allow us to approximate ( arbitrarily well ) the transformation : @xmath42 also , define @xmath43 . with these components ,
a network implementing @xmath44 is shown in figure [ u_k.fig ] .
, width=336,height=124 ] here we assumed that the algorithm works for a general family of states with classical descriptions @xmath33 . if we are only interested in producing a specific state @xmath16 the network can be simplified by removing the register containing @xmath33 and simplifying each @xmath37 to work only for that specific @xmath16 ( in the same way that one can simplify a circuit for adding variable inputs @xmath24 and @xmath45 to one that adds a fixed input @xmath46 to variable input @xmath45 ) .
the overall efficiency of our algorithm depends on how efficiently we can implement @xmath37 ; in other words , how efficiently we can compute the conditional probabilities @xmath47 or equivalently @xmath48 one example for which this is easy is the symmetric states .
the symmetric state @xmath49 is defined to be an equally - weighted superposition of the computational basis states @xmath50 that have hamming weight @xmath51 ( @xmath52 is the number of bits of @xmath24 that equal 1 )
. that is , @xmath53 the conditional probability @xmath54 is easily computed to be @xmath55 for @xmath56 , and to be @xmath57 for @xmath58 .
the hamming weight can be efficiently computed as shown in @xcite .
then we simply need to reversibly compute the @xmath40 satisfying @xmath59 + @xmath13 .
this technique will not allow us to generate efficiently all quantum states , but it will work for any family of states where for some reordering of the qubits we can efficiently compute the conditional probabilities @xmath62 .
99 richard cleve , artur ekert , chiara macchiavello , michele mosca .
`` quantum algorithms revisited '' proceedings of the royal society of london a , 454 , 339 - 354 , 1998 . on the quant - ph archive , report no .
huelga , s.f . ,
macchiavello , c. , pellizzari , t. , ekert , a. , plenio , m.b . and cirac j.i .
1997 , e - print quant - ph/9707014 .
p. kaye , m. mosca 2001 .
`` quantum networks for concentrating entanglement '' on the quant - ph archive , report no . 0101009 . | quantum protocols often require the generation of specific quantum states . we describe a quantum algorithm for generating any prescribed quantum state .
for an important subclass of states , including pure symmetric states , this algorithm is efficient . |
formation and evolutionary processes of stellar and planetary systems are expected to leave their imprint on the present - day systems .
one such imprint is the stellar obliquity , the angle between the stellar spin axis and the orbital angular momentum axis , also referred to as the spin - orbit angle . for star - planet systems
the measurement of this angle is a matter of intense study in recent years ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , primarily for hot jupiters gas - giant planets at short - period orbits .
some of the systems were found to be aligned , in a prograde orbit with spin - orbit angle close to zero , while others were found to be misaligned , including systems in retrograde motion where the spin - orbit angle is close to @xmath0 ( e.g. , * ? ? ?
* ; * ? ? ?
the growing sample and the wide range of spin - orbit angles measured for hot jupiters can be used for studying their orbital evolutionary history . for example , @xcite have noticed that hot stars , with an effective temperature above 6,250 k , tend to have a wide obliquity range , while cool stars tend to have low obliquities , mostly consistent with well aligned orbits .
this was confirmed by a study of a larger sample by @xcite and is consistent with the results of @xcite and @xcite who used different approaches .
those authors suggested that some mechanisms can cause the planetary orbit to attain large obliquity ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
then , tidal interaction with the host star ( e.g. , * ? ? ? * ) or magnetic braking ( e.g. , * ? ? ?
* ) act to realign the orbit .
since these processes are probably inefficient for hot stars , those systems might still retain their wide obliquity range .
so far spin - orbit alignment has been studied primarily through the rossiter - mclaughlin ( rm ) effect @xcite , originally suggested for stellar eclipsing binaries , and observed by monitoring the anomalous radial - velocity signal during eclipse , as the eclipsing star moves across the disc of the eclipsed star .
the rm effect is sensitive to the sky - projected component of the spin - orbit angle , and was successfully measured for many transiting planet systems ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , transiting brown dwarfs and low - mass star systems @xcite , and stellar binaries @xcite .
the line - of - sight component of the spin - orbit angle can be measured using asteroseismology @xcite , or the observed rotational broadening of spectral lines , _ if _ the host star radius and rotation period are known with sufficient precision ( * ? ? ?
* ; * ? ? ?
* see also @xcite ) .
however , these two methods require obtaining new data for each target , using valuable resources ( e.g. , large telescopes or _ kepler_short - cadence data ) .
other methods have been presented , based on stellar gravitational darkening @xcite , and the beaming effect ( photometric rm * ? ? ?
* ; * ? ? ?
an interesting approach was taken by @xcite and @xcite , who use the brief photometric signals during transit induced by the transiting object moving across spots located on the surface of the host object .
this is based on the fact that many stars show photometric modulations resulting from the combination of stellar rotation and non - uniform longitudinal spots distribution ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
when such a star displays transits by an orbiting planet , the transiting object might momentarily eclipse a stellar spot , inducing an increase in observed flux , if the surface brightness of the spot - covered area is lower than that of the non - spotted areas .
the derivation of the stellar obliquity requires identification of such ` spot - crossing ' events within a few transits , and estimate the spot and the planet phases within their motion over the stellar disc .
the method has since been applied to additional systems using _ high - speed _ _ kepler _ and corot data @xcite .
we present here another version of this approach that does not require such high - speed photometry .
instead , we use the fact that a spot - crossing event can induce measurable transit time variation ( ttv ; e.g. , * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , even for data that can not resolve the event itself .
our approach relies on the expected correlation between the induced ttv and the corresponding local photometric slope immediately outside the transit , presumably induced by the same spot .
detected correlation or anti - correlation between the ttvs and their local slope can in principle differentiate between prograde and retrograde rotation of the primary star in stellar binaries or star - planet systems .
we present here the basic concept and develop an analytical simplistic approximation for the induced ttvs and the photometric slope .
we also use the work of @xcite and @xcite , who developed a numerical tool soap - t , to simulate a planetary transit light curve which includes a spot - crossing event .
@xcite used soap - t to derive detailed transit light curves , and then fitted them with transit templates to obtain the expected ttvs , very similar to what is performed when deriving the ttvs from the _ kepler _ actual data ( e.g. , * ? ? ?
we show that our approximation yields ttvs with the same order of magnitude as the results of @xcite . using our approximation and the soap - t tool
we show that in some cases we can expect a negative ( positive ) correlation between the ttvs induced by spot crossing and the local photometric slopes at the transit timings for prograde ( retrograde ) motion of the planet .
we also discuss the limitations of this approach when applied to real data , showing that it can be applied only to a limited number of systems .
the paper is organized as follows .
section [ sec : principles ] outlines the basics of our approach , while section [ sec : ttv ] presents the analytical approximation for the induced ttv for different cases , and section [ sec : simulations ] compares our approximation with numerical simulations we performed and those of @xcite . in section [ sec : slopes ]
we derive the expected derivative of the stellar brightness at the time of transit , and in section [ sec : correlations ] display the expected correlation between the induced ttvs and the stellar photometric slopes .
finally , section [ sec : discussion ] discusses our results and the severe limitations of its applicability to real data .
the present paper is the first of three studies .
the next study ( holczer et al . , in preparation ) presents our analysis for the _ kepler _ planet candidates @xcite . in that paper
we show that indeed a few systems do show highly significant correlation between their derived ttvs and the local photometric derivatives , as predicted by this work .
a forthcoming paper will present our analysis of the _ kepler _ eclipsing stellar binaries @xcite .
to present our approach , we consider a transiting planet that crosses a stellar spot during its apparent motion over the stellar disc .
let us assume , for the sake of simplicity , that only one spot is present on the stellar disc and that the stellar rotation and orbital axes are parallel to each other .
this includes both prograde ( complete alignment , with obliquity of @xmath1 ) and retrograde ( obliquity of @xmath0 ) configurations .
the _ sign _ of the induced time shift depends on whether the spot - crossing event occurs in the first ( positive ttv ) or second ( negative ttv ) half of the transit , which is determined by the location of the spot on the stellar disc at the time of transit .
rotation as depicted in figures [ fig : prograde ] & [ fig : retrograde ] , the location of the spot over the stellar disc determines whether the star is becoming brighter or dimmer at the time of transit . when the spot is moving towards ( away from ) the center of the disc the stellar intensity is decreasing ( increasing ) , because of the aspect effect , which changes the effective area of the spot on the stellar visible disc the projected area of the spot onto the sky plane .
therefore , when the spot is on the disc edge its effective area is minimal . on the other hand
, the effective area reaches its maximum when the spot is at its closest position to the center of the visible disc , when the stellar surface is ( almost ) perpendicular to our line of sight .
another phenomenon , which also causes the star to become fainter when the spot moves towards the center of the disc is the limb - darkening effect , which is ignored at this point of the discussion .
now , when the stellar rotation and planetary motion have the same sense of rotation , the spot - crossing event in the first ( second ) half of the transit should always occur when the spot is moving towards ( away from ) the center of the disc .
therefore the _ signs _ of the induced ttv and the _ slope _ of the stellar brightness at the time of transit should be opposite .
this is depicted in figure [ fig : prograde ] for prograde motion . for retrograde motion ,
positive ttv should be associated with positive slope , as depicted in figure [ fig : retrograde ] .
therefore , we expect negative correlation between the derived ttvs and the corresponding stellar photometric slopes for a system with planetary prograde motion and positive correlation for a system with retrograde motion . in the next sections
we will show that this is indeed the case for a limited number of cases by deriving analytical approximations for the ttvs and the photometric derivatives and by numerical simulations for the ttvs .
to present the concept behind our method in a more quantitative way , figure [ fig : drawing ] shows a simplified schematic diagram of a transit light curve with a single spot - crossing event .
we neglect the transit ingress and egress finite duration of both the transit and the spot - crossing event .
we also assume that at the time of each transit there is only one circular spot on the stellar disc . in the figure we also neglect the limb - darkening photometric modulation , and will consider this effect later . in the figure ,
@xmath2 and @xmath3 are the depth of the transit and the amplitude of the photometric increase inside the transit due to the spot - crossing event , respectively , @xmath4 and @xmath5 are the duration of the transit and the spot - crossing event , respectively , and @xmath6 is the timing of the spot - crossing event relative to mid - transit time . from the figure one can see , using ` center - of - light ' formulation , that we expect the ttv induced by the spot - crossing event to be : @xmath7 this result is similar to equation ( 3 ) of @xcite after neglecting @xmath8 in the denominator .
we will adopt this approximation below . to derive the analytical simplistic model we first consider a case for which * the impact parameter of the spot and the planet are both equal to zero , namely that they both cross the center of the stellar disc , and * there is no limb darkening .
we lessen these two assumptions below .
we denote the location of the spot on the stellar disc by the angle @xmath9 , which is the angle between the observer and the spot , as seen from the stellar center . _ if _ the motion of the spot is equatorial , then @xmath9 is the longitude of the spot on the stellar visible hemisphere : @xmath10 where @xmath11 is the stellar angular velocity .
when the spot is on the stellar limb entering the visible hemisphere , @xmath9 gets the value of @xmath12 , and when the location of the spot is in the middle of its visible chord @xmath13 .
we denote the angle corresponding to the spot crossing by @xmath14 .
the sky - projected distance of the spot from the stellar center , as seen by the observer , is @xmath15 , where @xmath16 is the stellar radius .
the timing of the spot - crossing event , measured relative to the middle of the transit , is therefore @xmath17 in order to estimate the induced ttv , let us consider two extreme cases : a small spot , for which @xmath18 and a large spot , for which @xmath19 where @xmath20 and @xmath21 are the radii of the planet and the spot , respectively . for both cases
we introduce a darkness parameter , @xmath22 , which measures the surface brightness of the spot relative to the surface brightness of the star immediately outside the spot .
a completely dark spot would have @xmath23 , while @xmath24 close to unity means the spot is almost as bright as the unspotted stellar area .
for the small - spot approximation we can assume that the spot is completely covered by the planet during the spot - crossing event and therefore @xmath25 as noted above , the factor @xmath26 comes from the fact that the effective area of the spot is reduced by the aspect ratio , which is a function of the spot position on the visible stellar disc . for the large - spot approximation
the planet is fully contained in the spotted area during the spot - crossing event , and therefore we get @xmath27 here the factor @xmath26 comes from the fact that the time to cross the spot by the relatively small planet is reduced by the same aspect ratio .
we now approximate the ttv to be @xmath28 and the transit depth @xmath2 to be on the order of @xmath29 .
we therefore get for the small - spot approximation @xmath30 and for the large - spot approximation @xmath31 note that when @xmath32 , equation ( [ eq : timing_small1 ] ) @xmath33 equation ( [ eq_timing_large1 ] ) . to ease the discussion
we define @xmath34 as : @xmath35 using equation ( [ eq_t_sc ] ) we get : @xmath36 which is valid both for the small- and large - spot approximations .
the maximum observed ttv induced by the spot crossing is @xmath37 to include the stellar limb darkening effect in our model , we consider a quadratic limb - darkening law of @xmath38 , where @xmath39 is the scaled stellar surface brightness and @xmath40 and @xmath41 are the two limb - darkening coefficients , such that @xmath42 .
the induced ttv is proportional to @xmath3 , the increase of the stellar brightness during the spot crossing , which depends linearly on the stellar surface brightness @xmath39 , which is now a function of @xmath9 .
therefore we get @xmath43 note that because of the limb darkening the transit light curve does not have a rectangle shape , so our equation ( 1 ) should be modified .
nevertheless , as this analytical approach is aimed only to understand the features of the ttvs as a function of the spot - crossing phase , we neglect this effect that will affect all phases alike .
another extension of our simplistic model accounts for a non - zero impact parameter , @xmath44 .
note that the stellar rotation is , as before , orthogonal to our line of sight . in this extension of the simplistic model
, both planet and spot still have the same impact parameter , namely both move along the same chord on the stellar disc , a chord that does not go through the center of the disc .
therefore , the spot moves at a colatitude @xmath45 , with an impact parameter @xmath46 .
in such a case , the angle @xmath9 fulfill the relation @xmath47 where now @xmath48 is the longitude of the planet , and @xmath49 is when the planet crosses the projection of the stellar rotational axis .
the range of @xmath9 is now different : @xmath50 , and the timing of the spot crossing is @xmath51 where @xmath52 is the transit duration when @xmath53 .
a good approximation would be @xmath54 .
we now separate the discussion for the small and large spot approximations . for small spot ,
the duration of the spot - crossing event , @xmath5 , is still the same as for @xmath55 , but the transit duration is shorter by a factor of @xmath56 .
the flux increase depends on @xmath57 , as for @xmath55 .
we can therefore write @xmath58 combining these expressions we get @xmath59 for the large spot case , the duration of the spot - crossing event , @xmath5 , is now different , as the planet is crossing a spot which forms an ellipse on the stellar disc , whose axes are @xmath21 and @xmath60 .
one can show that the length of the planet s path on the spotted area is @xmath61 .
we therefore get @xmath62 and thus @xmath63 we can see that for a non - vanishing impact parameter there is a difference between the large and small planet cases , unlike in the basic model .
the difference is due to the @xmath64 term under the square sign in equation ( [ eq_ttv_large ] ) .
note that the approximation of the large spot is not valid for @xmath65 , where the projected area of the spot is small .
hence , we inserted into the calculation of the large - spot case a correction factor that turns the ttv expression to be similar to the small - spot one when @xmath66 .
this was done by multiplying the @xmath64 term with a fermi function that is approximately unity , except for @xmath66 , when the correction factor goes to zero . to further extend our simplistic model ,
we consider now a case for non - zero impact parameter _ and _ quadratic limb darkening together .
as before , we divide the discussion between the cases of small and large spot .
following equation ( [ eq_ttv_impact_small ] ) , but now multiplying it by the limb darkening brightness factor , we get for the small spot case : @xmath67 while for the large spot case , following equation ( [ eq_ttv_large ] ) , we get : @xmath68 the last case we consider is when the apparent planetary chord along the stellar disc goes through the center ( @xmath69 ) , but is inclined with the angle @xmath70 relative to the stellar equator .
we nevertheless assume that in some transits spot - crossing events happen , with spots that have different latitudes .
in such cases , @xmath6 is proportional to the distance of the spot - crossing event from the center of the disc , as in the basic model ( equation ( 3 ) ) .
similar considerations show that here also we get , as in equation ( 12 ) : @xmath71 which is true for small and large spot cases alike .
the extension for limb darkening also holds in this case . to visualize the expected ttvs derived by our analytical approximation for non - vanishing impact parameter cases , we plotted in figure [ fig : impact_parameter ] the calculated ttvs for different values of the impact parameter , with the large - spot approximation , using @xmath72 and @xmath73 values
. we chose a typical parameters for a transiting system a planet orbiting a star with solar radius in a 3 d orbit .
the duration of the transit ( mid - ingress to mid - egress ) is about 2.62 hours , a value on which we based our estimations .
one can see in the figure that the amplitude of the induced ttv is about 5 min .
the derived ttvs display almost linear slope as a function of the spot - crossing position , up to a maximum at distance of 0.60.85 stellar radii from the center of the stellar disc , and then a sharp drop to zero at the edge of the stellar disc .
as noted in the introduction , @xcite and @xcite developed a numerical tool soap - t , to simulate stellar photometric modulations induced by a rotating spot , including a planetary transit light curve which includes a spot - crossing event .
@xcite used soap - t to derive detailed transit light curves , and then fitted them with transit templates to obtain the expected ttvs , very similar to what is performed when deriving the ttvs from the _ kepler _ actual data ( e.g. , * ? ? ?
this is much more accurate derivation than that of the previous section , where we estimated the ttvs by the center - of - light approach .
it is therefore useful to compare the ttvs obtained by our analytical approximation with the ones derived with the soap - t numerical code and the transit fitting . to do that we perform in this section two comparisons .
first , we used ourselves the publicly available soap - t tool to produce transit light curves with spot - crossing events and fitted them with the @xcite codes to produce ttvs for a few cases and compare them with the analytical approximations .
second , we derive with our analytical center - of - light approach some ttvs for the cases derived by @xcite , and compare the results . in figure
[ fig : analytical_oshagh ] we plotted our analytical approximation for the same system as before a 3 d transiting planet orbiting a solar - like star .
we used limb darkening of @xmath74 and @xmath75 , @xmath76 and @xmath77 , a dark spot , with @xmath23 , and impact parameter of zero .
we can see from the figure that the maximum expected ttv based on our approximation is similar to the one obtained when simulating the spot - crossing event .
the obvious difference is the phase dependence while the analytical approximation has a smooth rise to the maximum , at phase of 0.65 , the simulated light curves yielded ttvs that are quite small for most phases , and rise sharply towards the maximum at phase 0.8 .
the reason for this difference comes from the different approaches of obtaining the ttv .
the approach that fits a model to the simulated light curve ignores sometimes the ` bump ' in the light curve caused by the spot - crossing event , yielding a small ttv , while the center - of - light model is , in fact , integrating over the whole transit light curve .
we will see this difference again and again . nevertheless , this difference does not change the result of this paper
the negative ( positive ) correlation for prograde ( retrograde ) motion , as will be shown below .
@xcite paper includes two figures that present their derived ttvs as a function of the orbital phase of the spot - crossing event .
we applied our analytical approximation to all cases included in @xcite figures , presented in the next two figures . in figure
[ fig : oshagh1 ] we plot results of our analytical approximation that corresponding to the six cases of @xcite figure 3 , where they have considered different spot and planet relative sizes , keeping the same limb darkening parameters .
we chose the same @xmath78 ( what they call f ) values 0.01 and 0.0025 , and the same @xmath79 values 0.05 , 0.1 , and 0.15 .
we used the same limb darkening coefficients of [ @xmath40,@xmath41 ] = [ 0.29,0.34 ] , and assumed a completely dark spot ( @xmath24 = 0 in our notation ) .
as before , the transit duration is set to be 2.62 hours . as in the previous figure , we see here that the maximum ttv is similar to the values obtained by @xcite , while the phase behavior of the two approaches is different , as explained above .
another comparison was done by constructing figure [ fig : oshagh2 ] and comparing it with figure 6 of @xcite , to study the effect of the limb darkening and spot darkness . here
again the amplitudes of the analytical approximation are similar to those of @xcite , while the phase dependence is different , like in our figure [ fig : oshagh1 ] .
we turn now to approximate the local photometric slope at the time of the transit , assuming as before that the stellar brightness is modulated by a single circular spot . for no limb darkening and null impact parameter we approximate the stellar flux , modulated by the spot as @xmath81 where @xmath82 is the observed amplitude of the photometric modulation .
this is so because the spot area on the stellar disc is reduced by the aspect ratio @xmath57 .
the _ derivative _ of the stellar photometric brightness is therefore @xmath83 the amplitude of the observed stellar photometric modulation is a function of the spot radius and darkness . to express this relation
we introduce the @xmath84 parameter , which accounts for the possibility that the spot crossed by the planet might not be the only spot that contributes to the stellar modulation with the observed phase .
therefore , @xmath85 measures the ratio of the area of the spot being crossed by the planet to the total neighboring spotted area that causes the photometric modulation _ with the same phase_. the total stellar modulation due to the spots , relative to the maximum stellar brightness , is @xmath86 in the case of limb darkening , the brightness of the spotted star takes the form @xmath87 as the photometry is modulated by the aspect ratio and the limb darkening at the spot s location .
the photometric derivative is then : @xmath88 the stellar photometry for non - vanishing impact parameter is expressed like in equation ( [ eq : stellar_photo ] ) , but now @xmath89 , and therefore the stellar photometry derivative is @xmath90 where @xmath91 is the time since the _ spot _ was in the middle of its trail , on the projection of the stellar spin ( see below ) on the stellar disc , and @xmath92 is the stellar rotation rate , as explained in section [ sec : obliquity ] . for non - vanishing impact parameter _ and _ stellar limb darkening the stellar photometry is @xmath93 and its derivative is @xmath94 when the obliquity of the system is non - vanishing , the spot moves on a chord orthogonal to the projection of the stellar rotational axis , at a colatitude @xmath95 , with @xmath96 .
the spot chord is different from that of the planet , which we assume goes through the center of the stellar disc .
because of the inclination of the transit chord , at the time of crossing @xmath97 where @xmath91 is the time since the _ spot _ was in the middle of its trail , on the projection of the stellar spin on the stellar disc , and @xmath92 is the stellar rotation rate .
therefore , the stellar photometric derivative is like equations ( [ eq : fdot_impact ] ) or ( [ eq : fdot_impact_limb ] ) , except for a @xmath98 factor .
note that when @xmath99 then @xmath100 , because the spot - crossing effect occurs near the photometric maximum , and therefore the correlation with the ttvs becomes difficult to detect .
we are now ready to consider the expected correlation between the ttvs induced by the spot - crossing events and the local slope of the stellar photometry at the time of the transit .
figure [ fig : ttvfdot ] displays our analytical approximation for the ttvs as a function of the photometric derivatives for a few cases .
the figure shows that the slope of the stellar brightness at the time of each transit and the corresponding induced ttv have _ opposite _ signs for prograde motion , and therefore we expect negative correlation between the two .
obviously , the slope and the induced ttv have the _ same _ sign for retrograde motion , because of the argumentation presented in section 2 and plotted in figures [ fig : prograde ] and [ fig : retrograde ] still holds , and therefore a positive correlation is expected in such a case .
figure [ fig : ttvfdot ] portraits how the ttvs derived by our analytical approximation depend on the photometric slope , but it does not show the real expected ttv , nor includes any observational noise , associated with every derived ttv and photometric derivative series . to see how these two affect the expected correlation we added normally distributed noise to both the simulated ttvs and the photometric derivatives , the results of which are plotted in figure [ fig : ttvvsdotanal ] for our analytically approximated ttvs , and in figure [ fig : ttvvsdotoshagh ] for ttvs derived by the soap - t tool . in both figures
we used the same fiducial system , but now with @xmath102 and @xmath103 = [ 0.29,0.34]$ ] .
the photometric derivative was scaled so that its maximum was unity .
we chose at random 500 phases to be plotted in the figures , and added randomly distributed normal noise to both the ttvs and the slope derivatives .
the noise r.m.s .
was equal to 50% of the maximum of the corresponding variable .
this amounts to 150 sec error on the ttvs and @xmath104 to the scaled slope .
the two figures show similar results there is a very clear anti - correlation between the induced ttvs and the photometric slopes at the transit timings , even when some small noise is added .
in fact , the noise covers up the fact that for some phases the dependence of the ttvs on the slope changes its sign , as we see in figure [ fig : ttvfdot ] . to estimate the expected effect of the noise on the measured correlation we ran extensive simulations , with different values of noise level and number of observed transits . for each choice of noise level , @xmath105 , @xmath106 and number of transits , @xmath107 , we chose @xmath107 random phases , derived their ttvs and photometric derivatives , added randomly distributed noise to both the ttvs and the stellar photometric slopes , and then derived the resulting ( anti-)correlation .
we repeated this simulation for 1000 times , with the same values of noise level and number of points .
we then derived the median and scatter of the sample of correlations obtained , which are plotted in figure [ fig : correlation_noise ] as a function of the noise level and @xmath107 .
we chose five values for @xmath105 and @xmath106 , each scaled as a fraction of the maximum of its corresponding variable .
the five noise - to - signal ratios we chose were [ 0 , 0.15 , 0.3 , 0.5 1 ] .
each choice characterizes both the noise added to the ttvs and to the photometric slopes . for @xmath107 we chose values of 50 , 100 , 500 , and 1000 . for short - period transiting planets _
kepler _ light curves could have on the order of 1000 transits , but 200400 was a more typical number .
all together we had @xmath108 sets of simulations , the results of which are plotted in figure [ fig : correlation_noise ] .
the expected value of the correlation depends on the noise level .
it goes from 0.9 for no noise down to 0.3 for a snr of unity .
the @xmath109 spread of the correlation depends on the noise level and the number of points .
it goes from 0.13 for @xmath110 and snr of unity down to 0.02 for @xmath111 and no noise .
the figure suggests that we can easily detect the correlation with snr of unity , if we can measure on the order of 500 ttvs and their corresponding photometric slopes .
we presented here a simple approach that can , in a few cases , use the derived ttvs of a transiting planet to distinguish between a prograde and a retrograde planetary motion with respect to the stellar rotation , assuming the ttvs are induced by spot - crossing events . using a simplistic analytical approximation we showed that those ttvs might have negative ( positive ) correlation with the local stellar photometric slopes at the transit timings for prograde ( retrograde ) motion .
we have shown that the correlation might be detected for different stellar limb darkening and different impact parameters .
furthermore , we obtained similar correlated ttvs when we used the soap - t tool to simulate transit light curves and derive the corresponding ttvs .
we have shown also that even if we include certain amount of noise , the correlation is still detectable .
can such a correlation surface above the observational noise ?
the expected amplitude of the ttv can be estimated by equations ( [ eq_factor2 ] ) and ( [ eq_timing_max ] ) .
for example , a system with @xmath112 , @xmath113 and transit duration of 3 h should show an induced ttv on the order of 5 min .
so , we can expect to observe the ( anti-)correlation between the ttvs and the photometric slopes only for systems with high enough signal - to - noise ratio that allows timing precision of the order of 5 min or better .
note that for a 3 d transiting planet in the _ kepler _ field we have at hand data for up to about 400 transits , enabling us to detect a correlation even if the noise is comparable with the signal . obviously , the approximation and simulation presented here are quite simplistic .
first , spotted stars probably have more than one spot . the spot eclipsed by the planet
might not be the one dominating the stellar flux modulation , and hence the local photometric slope at the time of transit might be very different from the expressions we developed here .
note , however , that in our simulation we allowed an error of the photometric slope that can be as large as the slope itself , and showed that even in such a case the correlation still can be detected .
second , spots have different stellar latitudes , so some transits might not have induced ttvs at all , contaminating the expected correlation . to deal with this problem
one might consider the correlation of only the highly significant ttvs , which could show the signal better .
third , the system obliquity can be very different from @xmath1 or @xmath114 , although most of the planets around cool stars , with a temperature below about 6000k , apparently are aligned with the stellar rotation @xcite .
we have shown that for systems with non - vanishing obliquity and null impact parameter the shape of the dependence of the ttv on the photometric slope is the same , although the obliquity might decrease or even eliminate the correlation , because many transits might not include a spot - crossing event at all .
note , however , that even for significant obliquity the correlation might still exist , assuming there will be enough induced ttvs , probably caused by spots with different latitudes . here
again one might ignore the non - significant ttvs when searching for a correlation .
fourth , the observed transiting system might have additional planets that induce dynamical ttvs , completely shadowing the ttvs caused by spot crossing events .
despite all these obstacles , the correlation studied here might be solid enough to show up for a few kois .
although our method can not give an accurate spin - orbit angle , but can instead only indicate the sign of the orientation of the planetary motion , the method might be useful nevertheless , as it uses _ kepler _ long - cadence data that is publicly available for all transiting planets . in the next paper ( holczer et al .
, in preparation ) we report on a search for correlation between the available ttvs and the corresponding local photometric slopes at the transit timings for all _ kepler _ kois , and indeed find five convincing cases with significant correlations .
the approach described here can in principle be applied to any eclipsing system , whether it is a transiting planet or a stellar binary . for a binary system ,
the induced observed minus calculated ( o - c ) eclipse timings can be estimated with the small - spot approximation , for which the planetary radius is that of the secondary .
we therefore expect the ttvs to be on the same order of magnitude as for transiting planets .
however , as eclipses in binaries are usually deeper and longer than the planetary transits , we expect the o - cs in eclipsing binaries to be more precise .
in fact , a negative correlation between the o - cs and the local photometric slopes was identified already for the stellar eclipsing binary in the kepler-47 circumbinary planet system @xcite .
the authors detected o - c on the order of 1 min in the timing of the primary eclipse , and used the derived linear trend to correct the eclipse timings .
the detection of a negative correlation for kepler-47 is consistent with a more detailed analysis of the spot - crossing events , also done by @xcite , which indicates a prograde motion .
the method presented here can be applied in the future to a large sample of systems monitored by current and future space missions , like k2 @xcite , tess @xcite , and plato @xcite , helping discovering , without additional observations , interesting systems that are worth following , and possibly find what are the conditions for alignment or misalignment of stellar rotations and orbital motions of planets and stellar binaries .
we are grateful to the referee for very helpful comments that helped us substantially improve the paper .
we are thankful to the authors of the soap - t tool that made it publicly available .
the research leading to these results has received funding from the european research council under the eu s seventh framework programme ( fp7/(2007 - 2013)/ erc grant agreement no
. 291352 ) .
t.m . also acknowledges support from the israel science foundation ( grant no.1423/11 ) and the israeli centers of research excellence ( i - core , grant no.1829/12 ) .
t.m . is grateful to the jesus serra foundation guest program and to hans deeg and rafaelo rebolo , that enabled his visit to the instituto de astrofsica de canarias , where the last stage of this research was completed .
this work was performed in part at the jet propulsion laboratory , under contract with the california institute of technology ( caltech ) funded by nasa through the sagan fellowship program executed by the nasa exoplanet science institute . | some transiting planets discovered by the _ kepler _ mission display transit timing variations ( ttvs ) induced by stellar spots that rotate on the visible hemisphere of their parent stars .
an induced ttv can be observed when a planet crosses a spot and modifies the shape of the transit light curve , even if the time resolution of the data does not allow to detect the crossing event itself .
we present an approach that can , in some cases , use the derived ttvs of a planet to distinguish between a prograde and a retrograde planetary motion with respect to the stellar rotation . + assuming a single spot darker than the stellar disc , spot crossing by the planet can induce measured positive ( negative ) ttv , if the crossing occurs in the first ( second ) half of the transit .
on the other hand , the motion of the spot towards ( away from ) the center of the stellar visible disc causes the stellar brightness to decrease ( increase ) .
therefore , for a planet with prograde motion , the induced ttv is positive when the local slope of the stellar flux at the time of transit is negative , and vice versa .
thus , we can expect to observe a negative ( positive ) correlation between the ttvs and the photometric slopes for prograde ( retrograde ) motion . using a simplistic analytical approximation , and also the publicly available soap - t tool to produce light curves of transits with spot - crossing events ,
we show for some cases how the induced ttvs depend on the local stellar photometric slopes at the transit timings . detecting this correlation in
_ kepler _ transiting systems with high enough signal - to - noise ratio can allow us to distinguish between prograde and retrograde planetary motions . in coming papers we present analyses of the kois and _ kepler _ eclipsing binaries , following the formalism developed here . |
the effect of reduced dimensionality on the electronic structure , especially on the ordered phases , has been one of the key issues in the condensed matter physics . for a one - dimension ( 1d ) system with short - range interactions ,
the thermal and quantum fluctuations prevent the development of a long - range order at finite temperatures@xcite , unless finite transverse coupling is present , i.e. the system becomes quasi-1d@xcite .
the scenario is exemplified with the discovery of the quasi-1d superconductor including the bechgaard salts ( tmtsf)@xmath6@xcite , @xmath7mo@xmath8se@xmath8 ( @xmath9=tl , in)@xcite , li@xmath10mo@xmath8o@xmath11@xcite , etc . in general , the bechgaard salts exhibits charge - density - wave or spin - density - wave instabilities and superconductivity emerges once the instability is suppressed by applying external pressure . for the molybdenum compounds , although the density - wave instabilities were absent in the superconducting samples , they were observed in the closely related @xmath12mo@xmath8se@xmath8 and @xmath13mo@xmath8o@xmath11 ( @xmath14=k , rb , etc ) .
therefore , a `` universal '' phase diagram@xcite suggests the superconductivity is in close proximity to the density - wave and most probably originated from non - phonon pairing mechanism .
however , the determination of the pairing symmetry for these compound was difficult due to the competition between several energetically close ordering parameters@xcite . recently , a new family of superconducting compounds @xmath12cr@xmath15as@xmath15 ( @xmath14=k , rb , cs ) was reported@xcite .
the crystal of the representing compound 233 ( @xmath16k ) consists of k atom separated ( cr@xmath15as@xmath15)@xmath17 nanotubes ( fig.[fig : geometry ] ) that are structurally very close to ( mo@xmath8se@xmath8)@xmath17 in the @xmath7mo@xmath8se@xmath8 ( mo-266 ) compounds .
however , unlike the mo-266 superconductors , 233 exhibits peculiar properties at both the normal and the superconducting states .
it shows a linear temperature dependent resistivity within a very wide temperature range from 7 to 300k at its normal state ; and most interestingly , its upper critical field ( @xmath18 ) steeply increases with respect to decreasing temperature , with extrapolation to 44.7 t at 0k , almost four times the pauli limit . in comparison ,
the upper critical field of tl@xmath19mo@xmath8se@xmath8 is 5.9 t and 0.47 t parallel and perpendicular to the quasi-1d direction@xcite , respectively .
such unusually high @xmath18 is beyond the explanation of multiband effect and spin - orbit scattering , and usually implies possible spin triplet pairing .
it is therefore quite essential to study the electronic structure of this compound , in particular its ground state , the spin - orbit coupling effect , as well as the possible spin fluctuations . in this paper
, we present our latest first - principles results on the 233 electronic structure .
we show that the ground state of 233 is paramagnetic .
the anti - symmetric spin - orbit coupling ( asoc ) effect is negligible far away from the fermi level , but sizable around it .
combining with the noncentral symmetric nature of the crystal , the sizable asoc effect around @xmath3 enables spin - triplet pairing mechanism@xcite .
the fermi surfaces of 233 consist of one 3d sheet and two quasi-1d sheets , and the cr-3d orbitals , in particular the d@xmath0 , d@xmath1 , and d@xmath2 orbitals dominate the electronic states around @xmath3 .
.total energies ( in mev / cr ) and cr magnetic moment ( in @xmath20 ) of considered magnetic states at different optimization levels .
the numbers given in / outside the parenthesis for the iaf states are the magnetic moments of cr@xmath21/cr@xmath22 atoms , respectively .
the numbers given in and outside the square brackets for the fm states correspond to @xmath23-axis and @xmath24-axis , respectively . [
cols="^,^,^,^,^,^,^,^,^ " , ] [ tab : state_energy ] we first examine the ground state of 233 .
fig.[fig : geometry ] shows the crystal structure of 233 .
the cr atoms form a sub - nanotube along the @xmath25-axis , surrounded by another sub - nanotube formed by as atoms .
the cr@xmath21-cr@xmath21 and cr@xmath22-cr@xmath22 bond lengths are respectively 2.61 and 2.69 , while the cr@xmath21-cr@xmath22 bond length is 2.61 .
therefore the six cr atoms in each unit cell ( as well as the six cr atoms from any two adjacent layers ) form a slightly distorted octahedron and become non - centrosymmetric .
since the cr sub - lattice is magnetically strongly frustruated , we considered several different possible magnetic states in our current study : the ferromagnetic state ( fm ) , the inter - layer antiferromagnetic state ( iaf)(fig .
[ fig : magiaf ] ) , the noncollinear antiferromagnetic state ( ncl)(fig .
[ fig : magncl ] ) , the noncollinear in - out state ( iop ) ( fig . [ fig : magiop ] ) , the noncollinear chiral - like state ( clk ) ( fig .
[ fig : magclk ] ) , and the nonmagnetic state ( nm ) .
the total energy calculation results are shown in tab . [
tab : state_energy ] . without structural optimization , all the magnetically stable states ( fm , iaf , iop and clk )
are energetically degenerate within the dft errorbar .
the presence of multiple nearly degenerate magnetic configurations indicates that the ground state is indeed paramagnetic phase .
it is important to notice that the converged iaf state yields -0.40 @xmath26 and 0.58 @xmath26 magnetic moment for cr@xmath21 and cr@xmath22 , respectively ; although it is initially constructed with zero total magnetization per unit cell
. such difference leads to a non - zero total magnetization in one unit cell , therefore the iaf state is indeed a ferrimagnetic state .
this phase can not be stabilized in dft unless the effective inter - cell magnetic interaction is ferromagnetic . with structural optimization
, the iaf state will eventually converge to fm state which is also energetically degenerate with nm state within the dft errorbar .
the above facts suggest that the system is close to magnetic instabilities but long range order can not be established .
therefore , we argue that the ground state of 233 is paramagnetic ( pm ) state .
we have also performed gga+@xmath27 calculations with @xmath27=1.0 and 2.0 ev , @xmath28=0.7 ev for the compound .
the inclusion of on - site @xmath27 strengthened the magnetic interactions , leading to more stable magnetic states .
as a result , all the non - collinear phases became energetically stable upon structural optimization even with a small @xmath27=1.0 ev .
nevertheless , these states are still energetically degenerate within lda errorbar , leaving the above conclusion unaltered .
however , the staggered moment per cr atom become @xmath29 30 times the experimental value with @xmath27=2.0 ev .
thus the gga+@xmath27 method significantly overestimates the magnetism in this system , suggesting that the gga functional better suits the current study . with closer examination of tab . [
tab : state_energy ] , it is important to notice that after full structural optimization ( where both lattice constants and atomic coordinates are allowed to relax ) , the resulting @xmath30 is 1.2% larger than the experimental value , but @xmath25 is 2% shorter .
similar situation was also observed in the studies of iron - based superconductors@xcite , and was due to the fact that the nm state in lda is only a very rough approximation to the actual pm state without considering any magnetic fluctuations . as the electronic structure close to @xmath3 seems to be sensitive to the structural changes ,
all the rest reported results are based on the experimental geometry unless specified , following the examples of iron - pnictides / chalcogenides .
figure [ fig : bsdos ] shows the band structure and density of states ( dos ) of 233 .
the band structure ( fig .
[ fig : bs ] ) shows flat bands at @xmath31 and @xmath32 around -0.93 ev , -0.86 ev , -0.64 ev , -0.56 ev , -0.37 ev , 0.51 ev and 0.95 ev relative to @xmath3 , forming quasi-1d van hove singularities in the dos plot at corresponding energies .
all bands including the flat bands mentioned above disperse significantly along @xmath33 , which clearly indicates the quasi-1d nature of the compound . from the band structure plot ,
it is apparent that there are three bands cross the fermi level , which are denoted as @xmath34 , @xmath35 and @xmath36 bands as shown in fig .
[ fig : bs ] , respectively .
the @xmath34 and @xmath35 bands are degenerate along @xmath5-a in the non - relativistic calculations , and they do not cross the fermi level in either the @xmath37 plane nor the @xmath32 plane , thus are likely to be quasi-1d bands . instead , the @xmath36 band crosses the fermi level not only along @xmath5-a , but also in the @xmath37 plane , therefore it is likely to be a 3d band .
the total dos at @xmath3 is calculated to be @xmath38=8.58 states/([email protected] . ) under non - relativistic calculations , with approximately 8% , 17% and 75% contribution from @xmath34 , @xmath35 and @xmath36 bands , respectively .
experimentally , the measured specific heat data yields @xmath36=70 mj/(k@xmath40mol ) , which corresponds to @xmath4129.7 states/([email protected] . ) , suggesting an electron mass renormalization factor @xmath42 , comparable to the one in the iron - pnictide case . the projected density of states ( pdos ) ( fig .
[ fig : dos ] ) shows that the cr-3d orbitals dominate the electronic states from @xmath3 - 1.0 ev to @xmath3 + 0.5 ev , and more importantly , the states from @xmath3 - 0.36 ev to @xmath3 + 0.26 ev are almost exclusively from the d@xmath0 , d@xmath1 , and d@xmath2 orbitals .
it is also instructive to notice that the five bands near @xmath3 has a band width of only @xmath40.6 ev at the @xmath37 plane in our current calculation , and would be @xmath40.2 ev if the electron mass renormalization factor is considered .
thus these bands are extremely narrow at @xmath37 plane , consistent with the quasi-1d nature of the compound .
we have also performed relativistic calculations to identify the spin - orbit coupling ( soc ) effect . since the magnitude of soc effect is proportional to the square root of the atomic number , the soc effect is initially expected to be negligible for this compound .
indeed , the soc effect on the @xmath34 band is negligible .
however , it is notable that the soc effect lifted the degeneracy between @xmath34 and @xmath35 bands at @xmath5 , as well as along @xmath5-a . more importantly ,
sizable ( @xmath460 mev ) asoc splitting can be identified on @xmath35 and @xmath36 bands around k close to @xmath3(insets of fig .
[ fig : bs ] ) . considering that the crystal structure lacks of an inversion center
, the presence of asoc splitting facilitates the spin - triplet pairing , which is a natural and reasonable explanation of its unusually high @xmath18 @xcite .
we plot the fermi surface sheets of 233 in fig .
[ fig : fs ] . the non - relativistic fermi surface of 233 consists of one 3d sheet formed by the @xmath36 band and two quasi-1d sheets formed by the @xmath34 and @xmath35 bands .
these sheets cut @xmath33 axis at @xmath43=@xmath44=0.30 @xmath45 and @xmath46=0.27 @xmath45 , respectively .
should the soc effect taken into consideration , the @xmath36 band is mostly affected and its fermi surface sheet splits into two ( fig . [
fig : fs]g ) .
nevertheless the two splitted @xmath36 sheets cut @xmath33 axis at the same point at @xmath46=0.25 @xmath45 . for the @xmath35 band ,
the soc effect causes the splitting of its fermi surface sheet , forming two adjacent quasi-1d sheets who cut @xmath33 axis at @xmath47=0.31 @xmath45 and @xmath48=0.33 @xmath45 , respectively .
the soc effect on @xmath34 band is negligible , thus the @xmath34 band and its fermi surface remain mostly unaltered and @xmath43 remains unchanged . we have also calculated the bare electron susceptibility @xmath49 ( fig .
[ fig : chi ] ) .
the real part of the susceptibility shows no prominent peak , consistent with our pm ground state result .
however , the imaginary part of the susceptibility shows extremely strong peaks at @xmath5 , which may be due to the large electron dos at the fermi level @xmath3 .
such a strong peak of imaginary part of susceptibility at @xmath5 is usually an indication of large fm spin fluctuation when finite on - site repulsion is taken into consideration .
considering the fairly large asoc effect , together with the non - central symmetric nature of the crystal structure , the large fm spin fluctuation may be the pairing mechanism and contribute to the spin - triplet pairing channel .
it should be noted that the strong peak of susceptibility imaginary part at @xmath5 is robust with respect to hole doping , as our rigid - band calculation of 0.2 hole doping shows the same feature .
our above studies are focused on calculations with experimental structure parameters .
however , it is important to point out that the number of fermi surface sheets and their respective dimensionality characters , the orbital characters of the bands close to @xmath3 , the soc effect on these bands , and the features of @xmath49 are insensitive with respect to the structural optimization ; therefore the physics is robust against the structural change ( supplementary information , fig .
si-2 , si-4 and si-6 ) .
nevertheless , the detailed band structure of 233 is quite sensitive to the structure relaxation . in particular
, the 3d @xmath36 band will submerge below @xmath3 at @xmath5 and the 3d @xmath36 sheet will be topologically different from the one shown in fig . [
fig : fs ] ( supplementary information , fig . si-3 and si-5 ) .
the quasi-1d @xmath34 and @xmath35 bands are much less affected , although they intersect with the @xmath33 axis at different @xmath50s . in conclusion
, we have performed first - principles calculations on the quasi-1d superconductor 233 .
the ground state of 233 is pm .
the electron states near the fermi level are dominated by the cr-3d orbitals , in particular the 3d@xmath0 , d@xmath1 , and d@xmath2 orbitals from @xmath3 - 0.5 ev to @xmath3 + 0.5 ev .
the electron dos at @xmath3 is less than 1/3 of the experimental value , suggesting intermediate electron correlation effect may in place .
three bands cross @xmath3 to form two quasi-1d sheets in consistent with its quasi-1d feature , as well as one 3d sheet who is strongly affected by the asoc splitting as large as 60 mev
. finally , the real part of the susceptibility is mostly featureless , consistent with our pm ground state result , but the imaginary part of the susceptibility shows large peaks at @xmath5 , indicating large ferromagnetic spin fluctuation exists in the compound . combined with the large asoc effect and the lack of inversion center in the crystal structure , the experimentally observed abnormally large @xmath18(0 ) may be due to a spin - triplet pairing mechanism .
first - principles calculations were done using vasp@xcite with projected augmented wave method@xcite and plane - wave .
perdew , burke and enzerhoff flavor of exchange - correlation functional@xcite was employed .
the energy cut - off was chosen to be 540 ev , and a 3@xmath513@xmath519 @xmath5-centered k - mesh was found sufficient to converge the calculations .
the k-3p , cr-3p and as-3d electrons are considered to be semi - core electrons in all calculations . the fermi surfaces were plotted by interpolating a five orbital tight - binding hamiltonian fitted to lda results using the maximally localized wannier function ( mlwf ) method@xcite . the same hamiltonian were also employed in the calculation of the bare electron susceptibility @xmath49 using : @xmath52\end{aligned}\ ] ] where @xmath53 and @xmath54 are the band energy ( measured from @xmath3 ) and occupation number of @xmath55 , respectively ; @xmath56 denotes the @xmath57 wannier orbital ; and @xmath58 is the number of the @xmath59 points used for the irreducible brillouin zone ( ibz ) integration .
the value of @xmath49 at @xmath60 was approximated by the value at @xmath61 with @xmath62 .
10 url # 1`#1`urlprefix[2]#2 [ 2][]#2 .
_ _ * * , ( ) .
_ _ * * , ( ) .
_ _ * * , ( ) . , , & _ _ * * , ( ) . , , , &
_ _ * * , ( ) .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _ _ * * , ( ) . &
_ _ , vol . ,
( , ) . .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ et al . _ . _
_ * * , ( ) . , &
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
& _ _ , chap .
( , ) . & .
_ _ * * , ( ) . .
_ _ * * , ( ) . & .
_ _ * * , ( ) . , & .
_ _ * * , ( ) . & .
_ _ * * , ( ) .
_ et al . _ . _
_ * * , ( ) .
the authors would like to thank fuchun zhang , zhuan xu , yi zhou and jianhui dai for inspiring discussions .
this work has been supported by the nsfc ( no . 11274006 and no .
11190023 ) , national basic research program ( no .
2014cb648400 and no .
2011cba00103 ) and the nsf of zhejiang province ( no .
lr12a04003 ) .
all calculations were performed at the high performance computing center of hangzhou normal university college of science .
h. j. performed most of the calculations ; g .- h . c. and c. c. were responsible for the data analysis and interpretation ; c. c. calculated the response functions and drafted the manuscript .
* competing financial interests * the authors declare no competing financial interests . | the electronic structure of quasi - one - dimensional superconductor 233 is studied through systematic first - principles calculations .
the ground state of 233 is paramagnetic .
close to the fermi level , the cr-3d@xmath0 , d@xmath1 , and d@xmath2 orbitals dominate the electronic states , and three bands cross @xmath3 to form one 3d fermi surface sheet and two quasi-1d sheets . the electronic dos at @xmath3
is less than 1/3 of the experimental value , indicating a large electron renormalization factor around @xmath3 . despite of the relatively small atomic numbers , the antisymmetric spin - orbit coupling splitting is sizable ( @xmath4 60 mev ) on the 3d fermi surface sheet as well as on one of the quasi-1d sheets .
finally , the imaginary part of bare electron susceptibility shows large peaks at @xmath5 , suggesting the presence of large ferromagnetic spin fluctuation in the compound . |
following @xcite , we further pursue the goal of an axiomatization of the category of categories based on the comprehensive factorization system ( cfs for short ) . here
, the emphasis is on the simultaneous consideration of both the cfs s @xmath4 and @xmath5 on @xmath6 , where @xmath7 ( @xmath8 ) is the class of final ( initial ) functors and @xmath9 ( @xmath10 ) is the class of discrete ( op)fibrations . rather than on the duality functor @xmath11
, we rest on the following reciprocal stability " property : _ the pullback of an initial ( final ) functor along a discrete ( op)fibration is itself initial ( final ) _ ( proposition [ 4 ] ) .
a balanced factorization category " is a finitely complete category @xmath0 with two factorization systems on it , which are reciprocally stable in the above sense and generate the same subcategory of internal sets " : @xmath12 .
it turns out that these simple axioms are remarkably powerful , allowing to develop within @xmath0 some important aspects of category theory in an effective and transparent way .
in particular , the reflection of an object @xmath13 in @xmath14 gives the internal set of components @xmath15 , while the reflections of a
point " @xmath16 in @xmath17 and @xmath18 are the slice @xmath19 and the coslice @xmath20 , with their maps to @xmath1 .
one can define two functors @xmath3 , which on @xmath13 give the underlying " categories @xmath21 and @xmath22 . furthermore , if @xmath23:= x\bs x \times_x x / y$ ] , the internal set @xmath24 $ ] enriches both @xmath25 and @xmath26 .
also composition in @xmath27 and @xmath28 can be enriched to a natural symmetrical " composition map @xmath29 , and it follows in particular that the underlying categories are dual .
in fact , the term balanced " is intended to reflect the fact that the objects of @xmath0 carry two dual underlying categories , coexisting on the same level .
it is the choice of one of the two factorization systems which determines one of them as preferred " : in fact , we will make only terminologically such a choice , by calling final " ( initial " ) the maps in @xmath7 ( @xmath8 ) and discrete fibrations " ( opfibrations " ) those in @xmath9 ( @xmath10 ) . for example
, we will see that the underlying functor @xmath3 determined by @xmath4 preserves discrete fibrations and opfibrations , final and initial points , slices and coslices and left and right adjunctible maps , while the other one reverses them .
the structure of balanced factorization category also supports the notion of balanced " cylinder and balanced " homotopy .
in particular , for any @xmath30 we have a cylinder ( see @xcite ) @xmath31^i\ar@<-.5ex>[r]_e & x[x , y ] \ar[r]^c & x(x , y ) } \ ] ] with @xmath32 , @xmath33 and @xmath34 , and pulling back along an element of @xmath35 ( corresponding to an arrow @xmath36 as defined above ) we get the arrow interval " @xmath37 $ ] : @xmath38^i\ar@<-.5ex>[r]_e & [ \alpha ] } \ ] ] which in @xmath6 is the category of factorizations of @xmath39 with the two factorizations through identities .
another consequence of the axioms is the following exponential law ( proposition [ 78 ] ) : if @xmath40 , @xmath41 and the exponential @xmath42 exists in @xmath2 , then it is in @xmath18 . as a consequence ,
if @xmath40 has a complement " ( see @xcite and @xcite ) : @xmath43 then it is valued in @xmath18 ( and conversely ) . in the balanced factorization category of posets ( see section [ ex ] ) , the internal sets are the truth values and the above enrichment gives the standard identification of posets with @xmath44-categories , while the complement operator becomes the usual ( classical ) one , relating upper - sets and lower - sets .
we here stress some properties of @xmath6 and of the comprehensive factorization systems @xmath4 and @xmath5 , on which we will model the abstraction of sections [ emcat ] and [ bal ] .
recall that given two functors @xmath45 and @xmath46 , the map ( or comma " ) construction @xmath47 yields as particular cases the slices @xmath48 or coslices @xmath49 , for any object @xmath16 .
more generally , @xmath50 and @xmath51 can be obtained as the pullbacks [ 1 ] @r=4pc@c=4pc p / x & x / x + p ^p & x @r=4pc@c=4pc xq & xx + q ^q & x in particular , the pullback [ 2 ] @r=4pc@c=4pc [ x , y ] & x / y + xx & x is the category of factorizations between @xmath52 and @xmath53 , which has as objects consecutive arrows @xmath54 and @xmath55 in @xmath1 , and as arrows the diagonals @xmath56 as below : [ 3 ] @r=3.5pc@c=3.5pc x ^ _ & z ^ + z ^^ & y applying the components functor @xmath57 we get @xmath58=x(x , y)$ ] .
indeed @xmath59 $ ] is the sum @xmath60 , \alpha : x\to y$ ] , where @xmath37 $ ] is the interval category " of factorizations of @xmath39 ; @xmath61 is an initial ( terminal ) object of @xmath37 $ ] iff @xmath62 ( @xmath63 ) is an isomorphism . for any functor @xmath64 and any @xmath65
, there is an obvious slice functor " of @xmath66 at @xmath52 : [ 3b ] e_f , x : x / xy / fx recall that a functor is a discrete fibration if it is orthogonal to the codomain functor @xmath67 , or , equivalently , if the object mapping of the slice functor @xmath68 is bijective , for any @xmath65 .
a functor @xmath45 is final if @xmath69 for any @xmath65 .
for example , an object @xmath70 is final iff it is terminal ( @xmath71 ) , the identity ( or any isomorphism ) @xmath72 is final ( @xmath73 , since @xmath20 has an initial object and so it is connected ) and @xmath74 is final iff @xmath1 is connected ( @xmath75 ) .
final functors and discrete fibrations form the left " comprehensive factorization system @xmath4 on @xmath6 , which is in fact the ( pre)factorization system generated by @xmath76 . since any @xmath4-factorization gives a reflection in @xmath17 ( see section [ emcat ] ) we have a left adjoint @xmath77 to the full inclusion of @xmath78 in @xmath79 .
the reflection formula " ( see @xcite and @xcite , and the references therein ) [ 3a ] ( p)x ( xp ) gives its value at @xmath52 .
dually , the domain functor @xmath80 generates the right " cfs @xmath5 : initial functors and discrete opfibrations ( this is the cfs originally considered in @xcite ) .
in particular , the reflection of an object @xmath16 in @xmath17 is the slice projection @xmath81 ( corresponding to the representable presheaf @xmath82 ) and the universal property of the reflection reduces to the ( discrete fibration version of the ) yoneda lemma .
so , a functor to @xmath1 is isomorphic in @xmath79 to the slice projection @xmath83 iff it is a discrete fibration whose domain has a terminal object over @xmath52 .
dually , if @xmath84 is the reflection in discrete opfibrations , @xmath85 in @xmath18 is the coslice projection .
on the other hand , the reflection of @xmath86 in @xmath87 gives the components of @xmath1 , that is , @xmath88 can be identified with the components functor @xmath89 .
the same of course holds for @xmath90 . [ 4 ] the pullback of an initial ( final ) functor along a discrete ( op)fibration is itself initial ( final )
. the two statements are clearly dual .
if in the diagram below @xmath91 is final , @xmath66 is a discrete opfibration , both squares are pullbacks and @xmath92 is a @xmath5-factorization of @xmath16 , [ 5 ] @r=4pc@c=4pc & xp & p _
p & q ^q + 1^i & xx ^x & x ^f & y we want to show that @xmath93 is connected .
since @xmath94 is a discrete opfibration whose domain has an initial object , it is a coslice of @xmath95 : @xmath96 ( @xmath97 is a @xmath5-factorization ) ; then the vertex of the pullback rectangle is @xmath98 , which is connected by hypothesis .
throughout the section , @xmath0 will be an @xmath4-category , that is a finitely complete category with a factorization system on it . for a brief survey on factorization systems and the associated bifibrations on @xmath0
we refer also to @xcite . we just recall that an @xmath4-factorization of a map @xmath64 in @xmath0 @xmath99^e\ar[dr]_f & z \ar[d]^m \\ & y } \ ] ] gives both a reflection of @xmath100 in @xmath101 ( with @xmath102 as reflection map ) and a coreflection of @xmath103 in @xmath104 ( with @xmath105 as coreflection map ) .
conversely , any such ( co)reflection map gives an @xmath4-factorization .
this can be restated in terms of arrow intervals : say that @xmath106 is a left ( right ) point of @xmath107 $ ] if @xmath108 ( @xmath109 ) is in @xmath7 ( @xmath9 ) .
then by orthogonality there is a unique map in @xmath107 $ ] from any left point to any right one , and an @xmath4-factorization of @xmath66 is a middle point " of the interval @xmath107 $ ] : it is final among left points and initial among right ones .
the above point of view may be reinforced by geometrical intuition ( as discussed at length in @xcite ) .
a map @xmath64 is a figure in @xmath95 of shape @xmath1 ( or a cofigure of @xmath1 in @xmath95 )
. a left point of @xmath107 $ ] factorizes the figure through an infinitesimal modification " of @xmath1 , while a right point factorizes it through a local aspect " of @xmath95 .
the middle point of the interval @xmath107 $ ] @xmath99^{e_f } & n(f ) \ar[r]^{m_f } & y } \ ] ] factorizes the figure through its neighborhood " @xmath110 .
if monic maps are concerned , a left point is an infinitesimal enlargement " of @xmath1 with respect to the embedding @xmath66 in @xmath95 , while a right point is an
open " part of @xmath95 which contains @xmath66 .
then the ( infinitesimal ) neighborhood @xmath110 is both the smallest open part of @xmath95 containing @xmath66 and the biggest infinitesimal enlargement of @xmath1 with respect to @xmath66 . in this vein , maps @xmath111 in @xmath9 are characterized as local homeomorphisms " ( if @xmath112 is a figure of @xmath1 , then @xmath113 gives the neighborhood @xmath114 ) ; on the other hand , maps @xmath115 in @xmath7 preserve global aspects " : if @xmath116 is a cofigure of @xmath95 and @xmath117 is the corresponding maximal infinitesimal enlargement of @xmath95 in @xmath118 , then @xmath119 is a maximal infinitesimal enlargement " @xmath120 of @xmath1 in @xmath118 .
however , our intuition of infinitesimal modification " should be broad enough to include , say , the dense inclusion of a part of a topological space in another one , and also the reflection map of a space in its set of components @xmath121 .
for example , if @xmath122 , components are one of the global aspects " preserved ( in the above sense ) by final functors .
we now define and analyze several @xmath4-concepts , which assume their usual meaning when @xmath122 with the left comprehensive factorization system ( see also @xcite , where the same topics are treated in a slightly different way , and other classical theorems are proved ) .
we say that a map @xmath33 is * final * , while a map @xmath123 is a * discrete fibration * ( df ) .
we assume that a canonical " factorization of any arrow in @xmath0 has been fixed , and denote by @xmath77 the corresponding left adjoint to the full inclusion @xmath124 .
if @xmath125 is a discrete fibration , we say that @xmath126 is an * internal set * or also a * @xmath0-set * ( or simply a set ) and we denote by @xmath127 the reflective full subcategory of @xmath0-sets ; @xmath14 is closed with respect to limits in @xmath0 , and so is itself finitely complete .
the reflection @xmath128 is the * components * functor .
so a final map @xmath129 to a set , that is a factorization of the terminal map @xmath86 , gives the set @xmath130 of components of @xmath1 , with the following universal property : [ 6 ] @r=3pc@c=3pc x ^e@/^1.5pc/[rr ] & s ^m @ .. >[r ] & s ^n + & 1 & that is , @xmath131 is initial among _ sets _ with a map from @xmath1 and , as remarked in section [ fact ] , it is also final among objects with a _ final _
map from @xmath1 .
in particular , a space @xmath1 is * connected * ( @xmath132 ) iff @xmath86 is final , iff any map to a set is constant .
the other important particular case is the reflection @xmath133 of a point " ( global element ) @xmath16 , which is obtained by factorizing the point itself : @xmath134^{e_x}\ar[dr]_x & x / x \ar[d]^{\down\,\ , x } \\ & x } \ ] ] we say that @xmath135 is the * slice * of @xmath1 at @xmath52 ( or the neighborhood " of @xmath52 , as in section [ fact ] and in @xcite ) .
the corresponding universal property is the following : [ 7 ] @r=3pc@c=3pc 1 ^e_x_x@/^1.5pc/[rr]^a & x / x ^ x @ .. >[r]^u & a ^m + & x & which in @xmath6 becomes the yoneda lemma : given a discrete fibration @xmath136 and an object @xmath137 of @xmath138 over @xmath1 , there is a unique factorization @xmath139 over @xmath1 through the universal element " @xmath140 .
note that @xmath141 is a discrete fibration too , and so displays @xmath19 as a slice @xmath142 of @xmath138 . given a map @xmath45 and a point @xmath52 of @xmath1 , a * cone * @xmath143 is a map over @xmath1 from @xmath144 to the slice @xmath19 : @xmath145^\gamma\ar[dr]_p & x / x \ar[d]^{\down\,\ , x } \\ & x } \ ] ] a cone @xmath146 is * colimiting *
if it is universal among cones with domain @xmath144 : [ 9 ] @r=3pc@c=3pc p ^_p@/^1.5pc/[rr]^ & x / x ^ x @ .. >[r]^u & x / y ^ y + & x & that is , a colimiting cone gives a reflection of @xmath144 in the full subcategory @xmath147 generated by the slices of @xmath1 ( the * underlying category * of @xmath1 ) .
if @xmath143 is final , it is an * absolute * colimiting cone . note that @xmath63 is indeed colimiting , since it gives a reflection of @xmath144 in @xmath17 . if @xmath148 , a cone @xmath149 is an * arrow * from @xmath52 to @xmath53 :
@xmath150^\alpha\ar[dr]_x & x / y \ar[d]^{\down\,\ , y } \\
& x } \ ] ] arrows can be composed via the kleisli construction : if @xmath149 , @xmath151 and @xmath152 is universally defined by the left hand diagram below , then @xmath153 is given by the right hand diagram : [ 10 ] @r=3pc@c=3pc 1 |e_y_y@/^1.5pc/[rr]^ & x / y ^y | & x / z ^z + & x & @r=3pc@c=3pc 1 |_x@/^1.5pc/[rr]^ & x / y ^y | & x / z ^z + & x & we so get a category which is clearly isomorphic to the underlying category @xmath27 , and which we will be identified with it whenever opportune .
note that the identities @xmath140 are fixed by the choice of a
canonical " @xmath4-factorization of points .
an arrow @xmath149 is colimiting iff it is absolutely so , iff it is an isomorphism in @xmath27 .
if @xmath154 is the reflection of @xmath45 in @xmath17 , the reflection map @xmath155 induces a bijection between cones @xmath156 and cones @xmath157 , which is easily seen to restrict to a bijection between the colimiting ones : [ 11 ] @r=3pc@c=3pc & & x / y@/^.5pc/[ddl]|<<<<<<<<<<y + p |e_p_p & n(p)^p | & x / x @ .. >[u]_u^x + & x & [ 12 ] if @xmath158 is a final map and @xmath45 , then a cone @xmath146 is colimiting , iff @xmath159 is colimiting .
since @xmath160 is a reflection map of @xmath161 in @xmath17 , by the above remark , @xmath162 is colimiting iff @xmath163 is colimiting , iff @xmath164 is colimiting . [ 13 ] @r=3pc@c=3pc t|e & p|e_p_p & n(p)^p | & x / x ^x + & & x & a map @xmath64 in @xmath0 acts on cones and arrows via an induced map between slices as follows ( see ( [ 3b ] ) ) . by factorizing @xmath165 as in the left hand diagram below
, we get a df over @xmath95 with a final point over @xmath166 .
then there is a ( unique ) isomorphism @xmath167 over @xmath95 which respects the selected final points . by composition , we get the right square below . its upper edge @xmath68 , which takes the final point of @xmath19 to that of @xmath168 , is the * arrow map * ( or * slice map * ) of @xmath66 ( at @xmath52 ) : [ 14 ] @r=4pc@c=3pc 1 ^e_x_x & x / x ^e^x & a ^m^~ & y / fx ^fx + & x ^f & y & @r=4pc@c=4pc x / x ^e_f , x^ x & y / fx ^fx + x ^f & y note that @xmath68 is uniquely characterized , among final maps @xmath169 , by the equation @xmath170 .
indeed , given another such @xmath171 , the dotted arrow induced among the @xmath4-factorizations is the identity , since it takes the universal point " @xmath172 to itself : [ 15 ] @r=4pc@c=4pc & & & y / fx^fx + 1^e_x & x / x ^e_f , x^e_f , x^ x & y / fx ^fx@ .. >[ur ] & + & x ^f & y & the arrow map of @xmath66 acts by composition on cones : it takes @xmath143 to its image cone " @xmath173 : [ 16 ] @r=4pc@c=4pc p ^_p@/^1.5pc/[rr]^f & x / x ^e_f , x^ x & y / fx ^fx + & x ^f & y which has the usual meaning in @xmath6 .
[ 17 ] an absolute colimiting cone is preserved by any map .
if in the diagram above @xmath63 is final , so is also @xmath174 .
in particular , @xmath66 induces a mapping between arrows @xmath175 of @xmath1 and arrows @xmath176 of @xmath95 , which preserves identities .
in fact , it is the arrow mapping of a functor @xmath177 : since in the left hand diagram below @xmath68 is final , there is a uniquely induced dotted arrow @xmath141 ( over @xmath95 ) which makes the upper square commute .
so , @xmath178 thus , @xmath179 and the right hand square commutes , showing that composition is preserved too . [
18 ] @r=3pc@c=3pc x / x @/_1.5pc/[dd ] _
^e_f , x & y / fx @ .. >[d]^u @/^1.5pc/[dd ] + x / x ^e_f , x & y / fx + x ^f & y @r=3pc@c=3pc x / x _ ^e_f , x & y / fx
^ + x / x^e_f , x & y / fx furthermore , @xmath180 ( since @xmath181 ) , [ 18a ] @r=4pc@c=4pc x / x ^e_f , x@/^1.5pc/[rr]^e_gf , x^ x & y / fx ^e_g , fx^fx & z / gfx ^gfx + x ^f & y ^g & z so that @xmath182 , and we get the * underlying functor * @xmath183 . [ 18b ] the underlying functor preserves the terminal object , sets , slices and discrete fibrations .
the fact that @xmath184 is the terminal category is immediate . for any final point @xmath185 in @xmath0 ,
the slice projection @xmath186 is an isomorphism , that is a terminal object in @xmath2 and so also in @xmath27 .
if @xmath64 is a df in @xmath0 , its arrow map at any @xmath52 is an isomorphism : [ 19 ] @r=4pc@c=4pc x / x ^~^ x & y / fx ^fx + x ^f & y thus , any arrow @xmath187 in @xmath188 has a unique lifing along @xmath189 to an arrow @xmath190 in @xmath27 , that is @xmath189 is a discrete fibration in @xmath6 .
the remaining statements now follow at once .
an example in section [ ex ] shows that connected objects ( and so also final maps ) are not preserved in general .
if the vertex of the pullback square below has a final point , we say that @xmath66 is * adjunctible * at @xmath53 : [ 20 ] @r=3pc@c=3pc 1 |e_x@/^1.5pc/[rr]^t & f / y ^p |q & y / y ^y + & x ^f & y and that the pair @xmath191 is a * universal arrow * from @xmath66 to @xmath53 .
in that case , @xmath192 is ( isomorphic to ) a slice @xmath19 of @xmath1 .
suppose that @xmath64 is adjunctible at any point of @xmath95 and denote by @xmath193 the point of @xmath1 corresponding to ( a choice of ) the final point @xmath102 , as in the left hand diagram below .
the right hand diagram ( whose squares are pullbacks ) shows that @xmath194 can be extended to a functor @xmath195 . ( note that we are not saying that there is a map @xmath196 in @xmath0 which gives @xmath197 as its underlying map . ) [ 21 ] @r=3pc@c=3pc 1 ^e_gy_gy & x / gy ^gy ^q & y / y ^y + & x ^f & y @r=3pc@c=3pc x / gy @/_1.5pc/[dd ] @ .. >[d ] & y / y @/^1.5pc/[dd ] + x / gy & y / y + x ^f & y [ 22 ] if @xmath64 is an adjunctible map , then there is an adjunction @xmath198 .
suppose that in the left hand diagram below the lower square is a pullback and that @xmath102 is the arrow map of @xmath66 .
since @xmath102 is final , any dotted arrow on the left induces a unique dotted arrow on the right which makes the upper square ( over @xmath95 ) commute .
since the lower square is a pullback , the converse also holds true .
the bijection is natural , as shown by the diagrams on the right .
[ 23 ] @r=3pc@c=3pc x / x @/_1.5pc/[dd ] @ .. >[d]^e & y / fx @ .. >[d ] @/^1.5pc/[dd ] + x / gy & y / y + x ^f & y @r=3pc@c=3pc x / x @/_1.5pc/[dd ] @ .. >[d]^e & y / fx @ .. >[d ] @/^1.5pc/[dd ] + x / gy & y / y + x / gy ^f & y / y @r=3pc@c=3pc x / x @/_1.5pc/[dd ] ^e & y / fx @/^1.5pc/[dd ] + x / x @ .. >[d]^e & y / fx @ .. >[d ] + x / gy ^f & y / y [ 24 ] the underlying functor preserves adjunctible maps . 1 .
diagram ( [ 20 ] ) shows that corollary [ 24 ] would be obvious if , along with slices , the underlying functor would preserve also pullbacks . up to now , the author does not know if this is true in general .
the adjunction of proposition [ 22 ] is nothing but the adjunction @xmath199 ( where @xmath200 gives , when @xmath122 , the left kan extension along @xmath66 ) , which for an adjunctible @xmath66 restricts to @xmath201 ( see @xcite ) .
3 . the diagrams below display the unit and the counit of adjunction @xmath202 : [ 25 ]
@r=3pc@c=3pc x / gfx ^q_gfx + & x / x @ .. >[ul]|-u|e_f , x^ x & y / fx ^fx + & x ^f & y @r=3pc@c=3pc & & y / fgy @ .. >[dl]|-v^fgy + x / gy |q^gy^e_f , gy & y / y ^y + x ^f & y in @xmath6 , @xmath141 is the functor @xmath203 , and the triangle @xmath204 reduces to the classical expression of the arrow mapping of a left adjoint functor via transposition and the unit @xmath205 .
similarly , @xmath206 is the functor @xmath207 , and the triangle @xmath208 becomes the classical expression of the transposition bijection of an adjunction via the arrow mapping of the left adjoint functor and the counit @xmath209 . [ 26 ] an adjunctible map @xmath64 preserves colimits : if @xmath146 is colimiting , then so is @xmath210 .
suppose that @xmath146 is colimiting .
we want to show that @xmath211 is colimiting as well , that is any cone @xmath212 factorizes uniquely through @xmath213 .
[ 27 ] @r=1.5pc@c=1.5pc & p ^f@ .. >[dd]|<<<<<u@/^1.5pc/[dddd]|<<<<<<<<<<<<<p_@/^2pc/[ddrr]^ + x / x @ .. >[dr]^u^<<<<<<<e_f , x & & y / fx @ .. >[dr]^u + & x / gy ^l & & y / y + + & x ^f & & y since the rectangle is a pullback , we have a universally induced cone @xmath214 , and since @xmath215 is colimiting also @xmath216 , which by adjunction ( @xmath68 is final ) corresponds to a unique @xmath217 , which is the desired unique factorization of @xmath63 through @xmath211 . note that we have followed the same steps of the most obvious classical proof , but at this abstract level we need not to verify or use any naturality condition .
if in the pullback below @xmath102 is colimiting we say that @xmath66 is * adequate * at @xmath53 : [ 28 ] @r=3pc@c=3pc f / y ^e & y / y ^y + x ^f & y while if @xmath102 is final we say that @xmath66 * dense*. note that in that case @xmath102 is an absolute colimiting cone and the diagram above displays at once dense maps as absolutely adequate " maps , as well as locally final " maps ( see @xcite for the case @xmath122 , where a different terminology is used ; see also section [ dense ] ) .
the density condition can be restated as the fact that the counit of the adjunction @xmath218 , that is the uniquely induced arrow @xmath206 on the left , is an isomorphism on the slices of @xmath95 : [ 30 ] @r=3pc@c=3pc & & _ f(f / y ) @ .. >[dl]|-v^m + f / y |q^p^e & y / y ^y + x ^f & y @r=3pc@c=3pc f / fx ^q_p + & x / x @ .. >[ul]|-u|e_f , x^ x & y / fx ^fx + & x ^f & y on the other hand , if the unit @xmath141 of the same adjunction , that is the uniquely induced arrow @xmath141 on the right , is an isomorphism on the slice @xmath19 , we say that @xmath66 is * fully faithful * at @xmath52 . note that if @xmath66 is fully faithful at @xmath52 , then @xmath66 is adjunctible at @xmath166 .
as shown in the previous section and in @xcite , category theory can in part be developed in any @xmath4-category .
however , in order to treat left - sided " and right - sided " categorical concepts simultaneously and to analyze their interplay , we need to consider _ two _ factorization systems , satisfying appropriate axioms : a * balanced factorization category * ( bfc ) is a finitely complete category @xmath0 with two factorization systems @xmath4 and @xmath5 on it , such that @xmath219 and satisfying the * reciprocal stability law * ( rsl ) : + the pullback of a map @xmath33 ( @xmath220 ) along a map @xmath221 ( @xmath123 ) is itself in @xmath7 ( @xmath8 ) .
while this definition aims at capturing some relevant features of @xmath6 , other intersesting instances of bfc are presented in setion [ ex ] .
note that any slice @xmath2 of a bfc is itself a weak " bfc : the requirement @xmath222 may not hold . as in section [ emcat ]
, we assume that canonical " @xmath4 and @xmath5-factorizations have been fixed for the arrows of @xmath0 , and denote by @xmath223 and @xmath224 the corresponding reflections .
for the concepts defined in section [ emcat ] , but referred to @xmath5 , we use the standard dual " terminology when it is available . otherwise , we distinguish the @xmath4-concepts from the @xmath5-ones , by qualifying them with the attributes left " and right " respectively .
so , we say that a map @xmath32 is * initial * and a map @xmath225 is a * discrete opfibration * ( dof ) .
the * coslice * of @xmath13 at @xmath16 is : @xmath134^{i_x}\ar[dr]_x & x\bs x \ar[d]^{\up\,\ , x } \\ & x } \ ] ] we denote by @xmath226 the class of bifibrations " .
of course , we also have a
right " underlying functor @xmath227 , with @xmath22 .
we begin by noting that , since @xmath12 ( that is , left " and right " @xmath0-sets coincide , and are bifibrations over @xmath228 ) , then also any ( op)fibration over a set is in fact a bifibration : @xmath229 .
furthermore , a map @xmath230 to a set is initial iff it is final : @xmath231 . in that case , @xmath232 , where @xmath233 is the component functor . in particular
, @xmath234 is the full subcategory of * connected * objects .
note also that a map @xmath235 between sets is final ( or initial ) iff it is an isomorphism ; so , if @xmath236 is a final ( or initial ) map from a set , then @xmath232 .
if @xmath237 is a reflection in @xmath14 , that is if @xmath238 is initial ( or equivalently final ) , for any element @xmath239 ( which is in @xmath240 ) , the pullback [ 31 ] @r=4pc@c=4pc [ s ] & x + 1^s & s displays @xmath241 $ ] as the component " @xmath241\inc x$ ] of @xmath1 at @xmath242 ; @xmath241 $ ] is connected by the rsl and is included in @xmath1 as a discrete bifibration .
( note that we are not saying that @xmath1 is the sum of its components , although this is the case under appropriate hypothesis . )
if furthermore @xmath243 is initial ( or final ) , then @xmath244 .
thus , @xmath245 is monomorphic and meets any component @xmath241 $ ] in a point , which is an initial point @xmath246 of @xmath241 $ ] : [ 32 ] @r=4pc@c=4pc 1^i_s & [ s ] & 1 ^s + s^i & x & s by a * balanced cylinder * , we mean a cylinder @xmath247^i\ar@<-.5ex>[r]_e & x \ar[r]^c & b } \ ] ] ( that is , a map with two sections ; see @xcite ) where the base inclusions @xmath245 and @xmath102 are initial and final respectively . as a consequence ,
@xmath238 is both initial and final , and if the base is a @xmath0-set @xmath126 , then @xmath232 and @xmath238 is a reflection in @xmath14 . among these discrete cylinders " ,
there are intervals " , that is balanced cylinders with a terminal base .
thus , intervals in @xmath0 are ( connected ) objects @xmath248 with an initial and a final selected point . in particular , any component of a cylinder with discrete base
is an interval : [ 33 ] @r=4pc@c=4pc 1@<.5ex>[r]^i_s@<-.5ex>[r]_e_s & [ s ] & 1 ^s + s@<.5ex>[r]^i@<-.5ex>[r]_e & x & s ( conversely , if @xmath7 and @xmath8 are closed with respect to products in @xmath249 , by multiplying any object @xmath250 with an interval , we get a cylinder with base @xmath251 . )
two parallel maps @xmath252 are * homotopic * if there is a balanced cylinder @xmath253^i\ar@<-.5ex>[r]_e & c \ar[r]^c & x } \ ] ] and a map ( omotopy " ) @xmath254 such that @xmath255 and @xmath256 ( see @xcite ) .
[ 34 ] the two base inclusions @xmath257 of a balanced cylinder are coequalized by any map @xmath258 toward a set .
since the retraction @xmath238 of the cylinder on its base is final , the proposition follows immediately from the fact that , by orthogonality , any map @xmath258 to a set factors uniquely through any final map @xmath259 : [ 35 ] @r=4pc@c=4pc c _ c & s + x @ .. >[ur ] & 1 [ 36 ] two homotopic maps @xmath252 are coequalized by any map @xmath260 toward a set . as in section [ cat ] , we denote by @xmath23 $ ] ( or simply @xmath59 $ ] ) the vertex of a ( fixed ) pullback [ 37 ] @r=4pc@c=4pc [ x , y ] ^q_x , y_p_x , y & x / y ^y + xx ^x & x we also denote by @xmath261\to x(x , y)$ ] the canonical " reflection map in @xmath14 .
thus @xmath262 ( or simply @xmath263 ) is the @xmath0-set @xmath264 $ ] . by further pulling back along the final ( initial ) point of the ( co)slice
, we get : [ 38 ] @r=4pc@c=4pc & s^e & 1 ^e_y + s^i & [ x , y ] & x / y + 1^i_x & xx & x where @xmath126 and @xmath265 are sets ( since the pullback of a discrete ( op)fibration is still a discrete ( op)fibration ) and @xmath245 ( @xmath102 ) is initial ( final ) by the rsl .
then @xmath266 and @xmath267 are both isomorphisms and in the above diagram we can replace @xmath126 and @xmath265 by @xmath263 , @xmath245 by @xmath268 and @xmath102 by @xmath269 .
so we have the following diagram of pullback squares : [ 39 ] @r=4pc@c=4pc & ( x , y)^e_x , y & 1 ^e_y + ( x , y)^i_x , y & [ x , y ] ^q_x , y^p_x , y & x / y ^y + 1^i_x & xx ^x & x and the balanced cylinder : [ 40 ] @c=3pc ( x , y)@<.5ex>[r]^i_x , y@<-.5ex>[r]_e_x , y & [ x , y ] ^c_x , y & ( x , y ) considering in particular the pullbacks [ 41 ] @r=4pc@c=4pc ( x , y)^q_x , yi_x , y & x / y + 1^x & x @r=4pc@c=4pc ( x , y)^p_x , ye_x , y & xx + 1^y & x we see that the elements of @xmath35 correspond bijectively to left " arrows @xmath36 and to right " arrows @xmath270 : [ 41b ] ( 1,x(x , y ) ) x(x , y ) (y , x ) then @xmath262 deserves to be called the internal hom - set of arrows from @xmath52 to @xmath53 " . ( note that by using @xmath59 $ ] and @xmath263 we have notationally made an arbitrary choice among one of two possible orders , causing an apperent asymmetry in the subsequent theory . ) from now on , also an element @xmath35 will be called an arrow from @xmath52 to @xmath53 " . composing it with @xmath271 ( @xmath272 ) we get the corresponding left ( right ) arrow ( which will be sometimes denoted with the same name ) . as in ( [ 33 ] ) , to any arrow @xmath35
there also corresponds an arrow interval " @xmath37 $ ] , the component at @xmath39 of the balanced cylinder ( [ 40 ] ) : [ 42 ] @r=4pc@c=4pc 1@<.5ex>[r]^i@<-.5ex>[r]_e^ & [ ] & 1 + ( x , y)@<.5ex>[r]^i_x , y@<-.5ex>[r]_e_x , y & [ x , y ] & ( x , y ) [ 43 ] two points @xmath30 are homotopic iff there is an arrow from @xmath52 to @xmath53 .
if @xmath149 is an arrow of @xmath1 , then @xmath37\to x$ ] gives a homotopy from @xmath52 to @xmath53 .
conversely , given a homotopy @xmath273^i\ar@<-.5ex>[r]_e & c \ar[r]^h & x } $ ] there is a ( unique ) arrow @xmath274 of @xmath275 , and so its image under @xmath109 is the desired arrow of @xmath1 .
the enriched composition " map @xmath276 is defined as @xmath277 , where @xmath278 is universally induced by the pullback : [ 44 ] @r=2pc@c=2pc & ( x , y)(y , z ) _
l^r@ .. >[d]^ & + & [ x , z ] ^c_x , z & + xx & ( x , z ) & x / z + & x & in which the maps @xmath279 and @xmath280 are the product projections followed by @xmath272 and @xmath281 , respectively .
[ 44a ] the map @xmath282 induces on the elements the composition mapping @xmath283 in @xmath27 . for any arrow @xmath284
we have a map @xmath285:[x , y]\to[x , z ] $ ] : [ 45 ] @r=2pc@c=2pc [ x , y ] @ .. >[dr]^[x,]^q_x , y_p_x , y & & x / y ^ + & [ x , z ] ^q_x , z^p_x , z & & x / z + + & xx & & x in the diagram below , the left dotted arrow ( induced by the universality of the lower pullback square ) and the right one ( @xmath286:(x , y)\to ( x , z)$ ] , induced by the universality of the reflection @xmath287\to ( x , y)$ ] ) are the same , since the horizontal edges of the rectangle are identities : [ 46 ] @r=4pc@c=4pc ( x , y ) @/_1.5pc/[dd]^i_x , y@ ..
>[d]^(x , ) & [ x , y]^c_x , y^[x , ] & ( x , y ) @ .. >[d]^(x , ) + ( x , z ) ^i_x , z & [ x , z]^c_x , z^p_x , z & ( x , z ) + 1 ^i_x & xx so we also have the commutative diagram [ 47 ] @r=4pc@c=4pc 1 ^ & ( x , y ) ^i_x , y^(x , ) & [ x , y]^q_x , y^[x , ] & x / y ^ + & ( x , z ) ^i_x , z & [ x , z]^q_x , z & x / z showing that @xmath288 acts on an arrow @xmath35 in the same way as @xmath152 acts on the corresponding left arrow @xmath289 , that is as @xmath290 . now observe that since @xmath291\circ i_{x , y}$ ] , and since @xmath292 and @xmath293 are trivially homotopic , also @xmath291\circ e_{x , y}$ ] . but
composing @xmath294\circ e_{x , y}:(x , y)\to[x , z]$ ] with the pullback projections @xmath295 and @xmath296 we get @xmath272 and the constant map through @xmath297 , so that the map @xmath294\circ e_{x , y}$ ] coincides with the composite @xmath298 ^ -\sim & ( x , y)\times 1 \ar[r]^-{\id\times\beta } & ( x , y)\times(y , z)\ar[r]^-{\mu ' } & [ x , z]}\ ] ] summing up , we have seen that : @xmath299 = c_{x , z}\circ[x,\beta]\circ i_{x , y } = c_{x , z}\circ[x,\beta]\circ e_{x , y } = \mu\circ ( \id\times\beta ) \circ \la \id,!\,\ra\ ] ] so , @xmath300 corresponds to @xmath301 , where @xmath302 is the composition in @xmath27 , proving the thesis .
dually " , for any @xmath149 there are maps @xmath303:[y , z]\to[x , z]$ ] and @xmath304 such that @xmath305 = c_{x , z}\circ[\alpha , z]\circ i_{y , z } = c_{x , z}\circ[\alpha , z]\circ e_{y , z } = \mu\circ ( \alpha\times\id ) \circ \la ! , \id \,\ra\ ] ]
so , @xmath306 corresponds to @xmath307 , where @xmath308 is the composition in @xmath309 .
[ 48 ] the two underlying categories are related by duality : @xmath310 . up to now
, the author has not been able to prove the associativity of the enriched composition , nor to find a counter - example . on the other hand ,
that the identity arrow @xmath311 [ 49 ] @r=2pc@c=3pc & 1 _ i_x^e_x@ .. >[d]^u_x & + & [ x , x ] ^c_x , x & + xx & ( x , x ) & x / x + & x & satisfies the unity laws for the enriched composition , follows from the previous section or can be easily checked directly . given a map @xmath64 in @xmath0 , we have an induced map @xmath312}:x[x , y]\to y[fx , fy]$ ] : [ 50 ] @r=2pc@c=2pc [ x , y ] @ .. >[dr]^f_[x , y]^q_x , y_p_x , y & & x / y ^e_f , y + & [ fx , fy ] ^<<<<<<<<<<q_fx , fy|<<<<<<<p_fx , fy & & y / fy + xx _ i_f , x & & x ^f + & fxy & & y in the diagram below the left dotted arrow ( induced by the universality of the lower pullback square ) and the right one ( @xmath313}:x(x , y)\to y(fx , fy)$ ] , induced by the universality of the reflection @xmath287\to ( x , y)$ ] ) are the same , since the horizontal edges of the rectangle are identities .
[ 51 ] @r=4pc@c=4pc ( x , y ) @/_2pc/[dd]^i_x , y@ ..
>[d]^f_x , y & [ x , y]^c_x , y^f_[x , y ] & ( x , y ) @ .. >[d]^f_x , y + ( fx , fy ) ^i_fx , fy & [ fx , fy]^c_fx , fy^p_fx , fy & ( fx , fy ) + 1 ^i_fx & fxy so we also have the commutative diagram [ 52 ] @r=4pc@c=4pc 1 ^ & ( x , y ) ^i_x , y^f_x , y & [ x , y]^q_x , y^f_[x , y ] & x / y ^e_f , y + & ( fx , fy ) ^i_fx , fy & [ fx , fy]^q_fx , fy & y / fy showing that @xmath314 acts on an arrow @xmath35 in the same way as the arrow map of @xmath66 acts on the corresponding left arrow @xmath289 .
so , @xmath315 enriches @xmath316 .
we also have the dual " commuting diagrams , with @xmath293 in place of @xmath292 and so on , showing that @xmath314 also enriches @xmath317 . [ 53 ] the two underlying functors are related by duality : @xmath318 . [ 55 ] the maps @xmath314 are functorial " with respect to enriched composition .
we want to show that the following diagram commutes : [ 56 ] @r=4pc@c=6pc ( x , y)(y , z ) ^-f_x , yf_y , z^ & ( fx , fy)(fy , fz ) ^ + ( x , z ) ^f_x , z & ( fx , fz ) since the lower rectangle below commutes [ 57 ] @r=4pc@c=6pc ( x , y)(y , z ) ^-f_x , yf_y , z^ & ( fx , fy)(fy , fz ) ^ + [ x , z ] ^f_[x , z]^c & [ fx , fz]^c + ( x , z ) ^f_x , z & ( fx , fz ) we have to show that the upper one also commutes , that is that composing with the pullback projections ( say , the second one @xmath319\to y / fz$ ] ) we get the same maps : [ 58 ] @r=4pc@c=4pc ( x , y)(y , z ) ^-^ & ( y , z ) |q_y , zi_y , z + [ x , z ] ^q_x , z^f_[x , z ] & x / z ^e_f , z + [ fx , fz ] ^q_fx , fz & y / fz @r=4pc@c=4pc ( x , y)(y , z ) ^-|f_x , yf_y , z & ( y , z ) ^f_y , z + ( fx , fy)(fy , fz ) ^-^ & ( fy , fz ) |q_fy , fzi_fy , fz + [ fx , fz ] ^q_fx , fz & y / fz indeed , this is the case because all the squares commute and the composites on the right are the two paths of the rectangle ( [ 52 ] )
. to any discrete fibration @xmath320 in @xmath0 there corresponds a presheaf @xmath321 on @xmath27 , as can be seen in several ways : 1 . as we have seen in proposition [ 18b ] ) , by applying @xmath183 to @xmath322 , we obtain a discrete fibration @xmath323 in @xmath6 , corresponding to a presheaf @xmath324 .
2 . via the inclusion @xmath325 ,
any object @xmath326 is interpreted " as a presheaf @xmath327 on @xmath27 .
3 . to any object @xmath16 of @xmath27
there corresponds the set @xmath328 of points of @xmath138 over @xmath52 , and to any arrow @xmath149 in @xmath27 there corresponds a mapping @xmath329 , whose value at @xmath330 over @xmath53 is given by the right hand diagram below ( where @xmath331 is defined by the left hand one ) : [ 59 ] @r=3pc@c=3pc 1 ^e_y_y@/^1.5pc/[rr]^a & x / y ^y ^a & a ^m + & x & @r=3pc@c=3pc 1 ^_x@/^1.5pc/[rr]^(m)a & x / y ^y ^a & a ^m + & x & thus , we have mappings @xmath332 , and we now show how they can be enriched as maps in @xmath14 .
( this enriching should not be confused with the @xmath333 of proposition [ 44a ] , associated to any map in @xmath0 . ) following and generalizing the results of section [ enrich ] , we define @xmath334 $ ] as the pullback [ 60 ] @r=4pc@c=4pc [ xm ] ^q_x , m_p_x , m & a ^m + xx ^x & x and @xmath335 $ ] , with reflection map @xmath336\to ( xm)$ ] .
then we also have the pullbacks : [ 61 ] @r=4pc@c=4pc ( xm)^i_x , m & [ xm ] ^q_x , m^p_x , m & a ^m + 1 ^i_x & xx ^x & x @r=4pc@c=4pc ( xm ) & a ^m + 1 ^x & x where we can assume that @xmath337 is the identity of the set @xmath338 , and which show that the elements of @xmath338 correspond bijectively to those of @xmath339 .
now we define ( see ( [ 44 ] ) ) @xmath340 as @xmath341 : [ 62 ] @r=2pc@c=2pc & ( x , y)(ym ) _
l^r@ .. >[d]^ & + & [ xm ] ^c_x , m & + xx & ( xm ) & a + & x & where @xmath278 is universally induced by the maps @xmath279 and @xmath280 , the product projections followed by @xmath272 and @xmath342 , respectively . to prove that this is the desired enrichment , for a fixed point @xmath330 over @xmath53 we consider the map @xmath343:[x , y]\to[xm]$ ] :
[ 63 ] @r=2pc@c=2pc [ x , y ] @ .. >[dr]^[x , a]^q_x , y_p_x , y & & x / y ^a + & [ xm ] ^q_x , m^p_x , m & & a ^m + + & xx & & x and define @xmath344:(x , y)\to(xm)$ ] , getting : [ 64 ] @r=4pc@c=4pc ( x , y ) @/_1.5pc/[dd]^i_x , y@ .. >[d]^(x , a ) & [ x , y]^c_x , y^[x , a ] & ( x , y ) @ .. >[d]^(x , a ) + ( xm ) ^i_x , m & [ xm]^c_x , m^p_x , m & ( xm ) + 1 ^i_x & xx @r=4pc@c=4pc ( x , y ) ^i_x , y^(x , a ) & [ x , y]^q_x , y^[x , a ] & x / y ^a + ( xm ) ^i_x , m & [ xm]^q_x , m & a the right hand diagram above shows that @xmath345 acts on an arrow @xmath35 in the same way as @xmath331 acts on the corresponding left arrow @xmath289 . furthermore we obtain that @xmath346 = c_{x , m}\circ[x , a]\circ i_{x , y } = c_{x , m}\circ[x , a]\circ e_{x , y } = \mu\circ(\id\times a)\circ\la\id,!\ra $ ] then , @xmath347 corresponds to @xmath348 , proving the thesis .
if @xmath349 is another discrete fibration over @xmath1 and @xmath350 a map over @xmath1 , we have a corresponding morphism of discrete fibrations @xmath351 in @xmath79 , or equivalently a natural transformation . on the other hand
, @xmath352 also induces maps @xmath353:[xm]\to[xn]$ ] : [ 65 ] @r=2pc@c=2pc [ xm ] @ .. >[dr]^[x,]^q_x , m_p_x , m & & a ^m^ + & [ xn ] ^q_x , n^p_x , n & & b ^n + + & xx & & x and one can check that the maps @xmath354:(xm)\to ( xn)$ ] define a morphism of left @xmath14-modules on @xmath1 , which enriches the natural transformation @xmath355 .
balanced category theory is category theory developed in a balanced factorization category @xmath0 , playing the role of @xmath6 with the comprehensive factorization systems .
we here present just a few aspects of it ; others still need to be analyzed . summarizing the results of section [ bal ]
, we have two underlying " functors : the left " one @xmath356 and the right " one @xmath357 , where @xmath358 and @xmath359 .
they are isomorphic up to the duality functor @xmath360 : [ 66 ] @r=1.5pc@c=6pc & ^(- ) + ^_ & + & by propositions [ 18b ] and [ 24 ] , both the underlying functors preserve the terminal object and sets ; the left one preserves final and initial points , discrete fibrations and opfibrations , slices and coslices and left and right adjunctible maps , while the right one reverses them .
this apparent asymmetry is only the effect of our naming of the arrows in @xmath7 , @xmath8 , @xmath9 and @xmath10 .
the duality of the underlying functors ( corollaries [ 48 ] and [ 53 ] ) is a consequence of the fact that for any object @xmath1 of @xmath0 we have an @xmath14-enriched weak category ( that is , the associativity of the composition @xmath361 may not hold ; see section [ weak ] ) , and for any map @xmath64 an @xmath14-enriched functor @xmath315 such that the following diagrams in @xmath362 commute : [ 67 ] @r=4pc@c=4pc x(z , y)(y , x)^ & ( 1,x(x , y)x(y , z))^~_~^(1 , ) & x(x , y)x(y , z)^ + x(z , x ) & ( 1,x(x , z))^~_~ & x(x , z ) [ 68 ] @r=4pc@c=4pc x(y , x)^f_y , x & ( 1,x(x , y))^~_~^(1,f_x , y ) & x(x , y)^f_x , y + y(fy , fx ) & ( 1,y(fx , fy))^~_~ & y(x , y ) then , denoting by @xmath363 the category of @xmath14-enriched weak categories and @xmath14-enriched functors , diagram ( [ 66 ] ) can be enriched as [ 69 ] @r=1.5pc@c=6pc & ^(- ) + ^_ & + & where @xmath364 , while @xmath365 .
furthermore , for any df @xmath136 , the corresponding presheaf on @xmath27 can also be enriched as a left @xmath14-module on the weak @xmath14-category @xmath1 , and the functor @xmath366 can be enriched to a functor @xmath367 ( and dually for @xmath368 ) . given @xmath369 , their * tensor product *
is the components set @xmath370 of their product over @xmath1 .
we so get a functor @xmath371 . in particular , @xmath372 and , if @xmath40 , @xmath373 .
if @xmath40 , @xmath41 and @xmath374 , @xmath375 gives the classical tensor product ( coend ) @xmath376 of the corresponding set functors @xmath377 and @xmath378 .
if @xmath46 is any map , and @xmath379 is a dof over @xmath1 , pulling back the @xmath4-factorization @xmath380 along @xmath381 we get : [ 72 ] @r=4pc@c=4pc nq ^ne_q|>>>>>>>>>n & nq & + d^n & q ^q^e_q & n(q)^q + & x & where @xmath382 denotes the product in @xmath2 .
since @xmath383 is a dof , by the rsl @xmath384 is a final map over @xmath1 ; thus its domain and codomain have the same reflection in @xmath17 : [ 73 ] if @xmath41 and @xmath385 , @xmath386 ; dually , if @xmath40 , @xmath387 similarly , given two maps @xmath45 and @xmath46 , we have the following symmetrical diagram : [ 74 ] @r=4pc@c=4pc & pq ^e & q ^e_q + pq ^i & pq & n(q ) ^q + p^i_p & n(p ) ^p & x since objects linked by initial or final maps have the same reflection in @xmath388 , we get the following inversion law " : [ 75 ] ( p , q ) ( p , q ) ( p , q ) if @xmath144 is a point @xmath16 , diagram ( [ 74 ] ) becomes : [ 76 ] @r=4pc@c=4pc & xq ^e & q ^e_q + ( xq ) ^i & [ xq ] & n(q ) ^q + 1^i_x & xx ^x & x and
the inversion law becomes the reflection formula " : [ 77 ] ( xq ) = ( xx , q ) ( xq ) = ( xx , q ) giving the enriched value of the reflection @xmath389 at @xmath52 .
if the functor [ 77a ] ^x m ( -x , m ) reflects isomorphisms , then the formula @xmath390 determines @xmath391 up to isomorphisms .
in particular , in this case , @xmath91 is final iff the inverse image @xmath392 of any slice of @xmath1 is connected : @xmath393 .
if @xmath122 , we find again the classical " formula ( [ 3a ] ) : @xmath394 which gives the presheaf corresponding to the discrete fibration @xmath391 . in section [ emcat ] , we have already discussed dense maps , from a left - sided " perspective .
but density is clearly a balanced " concept .
indeed , the following is among the characterizations of dense functors in @xmath6 given in @xcite : [ 70 ] if @xmath64 is dense , then for any arrow @xmath39 of @xmath95 the object @xmath395 , obtained by pulling back the arrow interval @xmath37 $ ] along @xmath66 , is connected .
if the functor ( [ 77a ] ) reflects isomorphisms , the reverse implication also holds .
the pullback square factors through the two pullback rectangles below : [ 71 ] @r=2pc@c=2pc //f ^e & & [ ] ^m + & f / y ^<<<<<<<<<<e & & y / y + x ^f & & y if @xmath66 is dense , the map @xmath102 is final by definition ; @xmath396 is a discrete opfibration , because it is the composite of the component inclusion @xmath37\to[x , y]$ ] and the pullback projection @xmath397\to y / y$ ] , which are both discrete opfibrations ( see diagrams ( [ 37 ] ) and ( [ 42 ] ) ) .
then , by the rsl also @xmath398 is final , and the connectedness of @xmath37 $ ] implies that of @xmath395 . for the second part , note that @xmath37 $ ] , having an initial point , is a coslice of @xmath399 : [ 71a ] [ ] ( y / y ) thus , @xmath395 is the pullback @xmath400 , and the fact that it is connected for any @xmath39 implies , by the hypothesis , the finality of @xmath398 .
till now , we have not considered any of the exponentiability properties of @xmath6 . in this subsection
, we assume that for any @xmath13 , the discrete fibrations and opfibrations over @xmath1 are exponentiable in @xmath2 .
[ 78 ] if @xmath40 and @xmath41 then @xmath401 and , dually , @xmath402 .
if @xmath381 is a dof over @xmath1 and @xmath46 is any object in @xmath2 , by proposition [ 73 ] , @xmath403 is a final map over @xmath1 ( recall that @xmath404 is the reflecton map of @xmath385 in @xmath17 , that is the unit @xmath405 of the adjunction @xmath406 ) .
then it is orthogonal to any df over @xmath1 , in particular to @xmath105 . in turn
, this implies that any unit @xmath407 is orthogonal to @xmath408 , since in general if @xmath409 , an arrow @xmath410 in @xmath240 is orthogonal to the object @xmath411 iff @xmath412 is orthogonal to @xmath138 in @xmath413 : [ 79 ] @r=4pc@c=4pc b ^h_f & a + b @ .. >[ur]^u _ * & 1 @r=4pc@c=4pc lb ^h^*_lf & ra + lb @ .. >[ur]^u & 1 indeed , in one direction the right hand diagram above gives a unique @xmath141 , whose transpose @xmath414 makes the left hand one commute : @xmath415 .
the converse implication is equally straightforward .
then , taking @xmath416 we see that @xmath417 is a reflection map with a retraction , and so @xmath401 ( see @xcite or also @xcite ) .
any @xmath40 has a complement " [ 79a ] ( m,-)m :
s(sx)^m where @xmath418 is the constant bifibration over @xmath1 with value @xmath126 . by proposition
[ 78 ] , @xmath419 takes values in @xmath18 , giving a broad generalization of the fact that the ( classical ) complement of a lower - set of a poset @xmath1 is an upper - set .
( these items are trated at length in @xcite and @xcite .
there , we conversely used the exponential law to prove the inversion law in @xmath6 ; however , the present one seems by far the most natural path to follow , also in the original case @xmath122 . )
an object @xmath13 is * codiscrete * if all its points are both initial and final ( that is , all the ( co)slice projections are isomorphisms ) .
an object @xmath13 is * grupoidal * if all its ( co)slices are codiscrete .
( note that codiscrete objects and sets are grupoidal . ) since these concepts are autodual and the underlying functors preserve ( or dualize ) slices , it is immediate to see that they preserve both codiscrete and grupoidal objects ( which in @xmath6 have the usual meaning ) . * if @xmath0 is an @xmath4-category such that @xmath7 is @xmath9-stable ( or , equivalently , satisfying the frobenius law ; see @xcite ) , then @xmath0 is a bfc with @xmath420 . * in particular
, for any lex category @xmath0 we have the discrete " bfc @xmath421 and the codiscrete " bfc @xmath422 on @xmath0 . in @xmath423 , @xmath424 and @xmath425 .
conversely in @xmath426 , @xmath427 and @xmath428 . in @xmath423
all object are discrete : @xmath429 , while @xmath228 is the only connected object .
conversely , in @xmath426 all objects are connected , while @xmath228 is the only set .
so , for any @xmath430 ( @xmath431 ) @xmath432 ( @xmath74 ) is the reflection map of @xmath1 in sets and , for any @xmath16 , @xmath433 ( @xmath434 ) .
then , for any @xmath430 ( @xmath431 ) , @xmath27 is the ( co)discrete category on @xmath435 .
+ furthermore , if @xmath430 and @xmath30 , then @xmath23 $ ] is pointless if @xmath436 , while @xmath437=1 $ ] .
so the same is true for @xmath262 . on the other hand , if @xmath431 , then @xmath23=x$ ] and @xmath438 , for any @xmath30
indeed , for any arrow @xmath439 , the arrow interval @xmath440 $ ] is simply @xmath30 .
* on the category @xmath441 of posets , we can consider the bfc in which @xmath123 ( @xmath10 ) iff it is , up to isomorphisms , a lower - set ( upper - set ) inclusion , and @xmath442 is in @xmath7 ( @xmath8 ) iff it is a cofinal ( coinitial ) mapping in the classical sense : for any @xmath65 , there is @xmath443 such that @xmath444 ( @xmath445 ) .
+ the category of internal sets is @xmath446 , and the component functor @xmath447 reduces to the non - empty " predicate . for any poset @xmath1 and any point @xmath65 ,
the slice @xmath19 ( @xmath20 ) is the principal lower - set ( upper - set ) generated by @xmath52 . as in @xmath6 ,
@xmath27 is isomorphic to @xmath1 itself .
this example also shows that the underlying functor may not preserve final maps : a poset @xmath1 is internally " connected iff it is not empty , while @xmath448 may well be not connected in @xmath6 .
+ given a map @xmath45 , a colimiting cone @xmath146 simply indicates that @xmath52 is the sup of the set of points @xmath449 .
the colimit is absolute iff such a sup is in fact a maximum .
a map @xmath64 is adequate iff any @xmath450 is the sup of the @xmath166 which are less than or equal to @xmath53 .
+ furthermore , given @xmath30 , @xmath59\subseteq x$ ] has its usual meaning : it is the interval of the points @xmath451 such that @xmath452 .
so , @xmath262 is the truth value of the predicate @xmath453 , and the enriching of @xmath27 in internal sets gives the usual identification of posets with categories enriched over @xmath44 .
+ by restricting to discrete posets , we get the bfc associated to the epi - mono factorization system on @xmath362 . *
let @xmath454 be the category of reflexive graphs , with the factorization systems @xmath4 and @xmath5 generated by @xmath67 and @xmath80 respectively , where @xmath44 is the arrow graph ( see @xcite and @xcite ) . in this case @xmath27 is the free category on @xmath1 , and the underlying functor is the free category functor @xmath455 .
+ furthermore , if @xmath456 and @xmath30 , then @xmath23 $ ] is the graph which has as objects the pairs of consecutive paths @xmath457 , with @xmath458 and @xmath459 , while there is an arrow @xmath460 over @xmath461 iff @xmath462 and @xmath463 . in @xmath454 , @xmath464 and the components of @xmath23
$ ] are the paths @xmath36 .
thus , the enriching @xmath262 of @xmath25 is in fact an isomorphism .
there are at least three well - estabilished abstractions ( or generalizations ) of category theory : enriched categories , internal categories and 2-category theory .
each of them is best suited to enlighten certain of its aspects and to capture new instances ; for example , monads and adjunctions ( via the triangular identities ) are surely 2-categorical concepts , while enriched categories subsume many important structures and support quantifications . here , we have based our abstraction on final and initial functors and discrete fibrations and opfibrations , whose decisive relations are encoded in the concept of balanced factorization category ; in this context , natural transformations and/or exponentials are no more basic notions .
rather , we have seen how the universal exactness properties of @xmath0 and those depending on the two factorization systems , complement each other in an harmonious way ( in particular , pullbacks and the reflection in discrete ( op)fibrations ) to give an effective and natural tool for proving categorical facts . while not so rich of significant instances different from @xmath6 as other theories , we hope to have shown that balanced category theory offers a good perspective on several basic categorical concepts and properties , helping to distinguish the trivial " ones ( that is , those which depend only on the bfc structure of @xmath6 ) from those regarding peculiar aspects of @xmath6 ( such as colimits , lextensivity , power objects and the arrow object ) .
the latter have been partly considered in @xcite and deserve further study .
* _ simple _ : a bfc is a lex category with two reciprocally stable factorization systems generating the same ( internal ) sets ; * _ expressive _ : many categorical concepts can be naturally defined in any bfc ; * _ effective _ : simple universal properties guide and almost force " the proving of categorical properties relative to these concepts ; * _ symmetrical _ ( or balanced " ) : the category @xmath6 in itself does not allow to distinguish an object from its dual ; this is fully reflected in balanced category theory ; * _ self - founded _ : it is largely enriched on its own internal sets , providing in a sense its own foundation ( see for example @xcite and @xcite ) . | some aspects of basic category theory are developed in a finitely complete category @xmath0 , endowed with two factorization systems which determine the same discrete objects and are linked by a simple reciprocal stability law .
resting on this axiomatization of final and initial functors and discrete ( op)fibrations , concepts such as components , slices and coslices , colimits and limits , left and right adjunctible maps , dense maps and arrow intervals , can be naturally defined in @xmath0 , and several classical properties concerning them can be effectively proved . for any object @xmath1 of @xmath0 , by restricting @xmath2 to the slices or to the coslices of @xmath1 , two dual underlying categories " are obtained .
these can be enriched over internal sets ( discrete objects ) of @xmath0 : internal hom - sets are given by the components of the pullback of the corresponding slice and coslice of @xmath1 .
the construction extends to give functors @xmath3 , which preserve ( or reverse ) slices and adjunctible maps and which can be enriched over internal sets too . |
a fundamental ingredient of almost every biological process is molecular recognition , which is widely observed in interaction systems like protein - protein , receptor - ligand , antigen - antibody , dna - protein , rna - ribosome , etc@xcite .
understanding the basis for molecular recognition requires the full characterization of binding process which involves noncovalent bonding effect and biomolecular thermodynamics . among all the factors ,
biomolecular conformation energy is of great importance@xcite .
it is found that changes in protein conformational entropy can contribute significantly to the free energy of protein - ligand association@xcite .
internal dynamics of the protein calmodulin varies significantly on binding a variety of target domains .
the apparent change in the corresponding conformational entropy is linearly related to the change in the overall binding entropy .
also , conformational entropy of protein side - chain is a major effect in the energetics of folding .
the entropy of heterogeneous random coil or denatured proteins is significantly higher than that of the folded native state tertiary structure . in particular ,
the conformational entropy of the amino acid side chains in a protein is thought to be a major contributor to the energetic stabilization of the denatured state and thus a barrier to protein folding .
the reduction in the number of accessible main chain and side - chain conformation when a protein folds into a compact globule yields an unfavourable entropic effect .
this reduction in conformational entropy counters the hydrophobic effect favoring the folded state and in part explains the marginal stability of most globular proteins . even thought biomolecular conformation energy is a key property to understand a wide variety of physical , chemical , and biochemical phenomena , its evaluation or calculation is very challenging both experimentally and theoretically . only recently ,
nmr relaxation methods for characterizing thermal motions on the picosecond - nanosecond ( ps - ns ) timescale is developed and the resulting order parameters is used as a proxy for conformational entropy evaluation@xcite .
atomic force microscopy ( afm ) for unfolding has great potential in measuring the backbone conformational entropy for protein folding@xcite .
neutron spectroscopy is also used to elucidate the role of conformational entropy upon thermal unfolding by observing the picosecond motions , which are dominated by side - chain reorientation and segmental movements of flexible polypeptide backbone regions@xcite .
the computational estimation of conformation energy is a long - standing problem and one of challenges in computational chemistry@xcite .
generally , the calculation of conformational entropy requires a fully understanding of biomolecular configuration spaces .
various simulation techniques , including harmonic analysis , molecular dynamics ( md ) , monte carlo simulation , normal mode analysis and so on , are all employed to explore the configuration spaces@xcite .
traditionally , the simulation data is processed by the quasiharmonic analyses@xcite .
more specifically , a special variance matrix for the biomolecular conformations can be defined and analyzed with principal component analysis ( pca ) .
conformation entropies are then evaluated from the eigenvalue and eigenvector information from pca . however , for macromolecules , quasiharmonic analyse can be computationally demanding . and the employment of the cartesian coordinates representation makes it very inefficient for bond rotation characterization@xcite . moreover , it is found that , for biomolecular systems with multiple occupied energy wells , quasiharmonic approximation tends to overestimate their configurational entropy , and with cartesian coordinates , the errors tend to be magnified@xcite .
characterization of biomolecular conformations by some structural parameters has been proposed to estimate the conformational entropy@xcite .
particularly , backbone dihedral angles and side chain rotamers are widely used structural measurements for protein conformational entropy evaluation .
the entropy value depends on the probability of the occupancy of the particular structure . by the assumption that amino acids distributions in the native state of proteins are comparable to that found in the denatured state , stites and pranata@xcite proposed a way to evaluate the relative conformation entropies for different amino acids .
they analyze the preferred distribution of amino acid residues by systematically studying about twelve thousand residues from sixty - one nonhomologous and high - resolution protein crystal structures .
a ramachandran plot of @xmath0 angles for various amino acid residues are obtained and the conformation entropies are evaluated through a discretization of angle distribution by an uniformed grid size .
the dihedral angle and sid - chain rotamer based parameterization has been widely used in biomolecular conformation entropy estimation@xcite .
the above - mentioned dihedral - angle based entropy evaluation method usually involves a discretization of angle distributions and it is found that the calculated entropy values are sensitive to the grid size@xcite .
it is simply true that when the grid is very fine or dense , angles with similar values can be classified into different categories .
in contrast , when the grid is very coarse , even angles with large variation can still be classified into a same category .
therefore , dramatically different entropy values can be obtained from the same data under different discretization .
an example can be found in figure [ fig : entropy_problem ] .
however , it has also been pointed out that correlation coefficients between entropies computed from different meshes are very high , suggesting that mesh - related bias does not systematically alter relative entropy values@xcite . but this consistence is highly related to the studied systems . simply speaking , for each ramachandran plot ,
only the same type of amino acide is considered thus the angle is highly concentrated in a particulary area instead of scattering over a wide range .
researchers have realized that a lack of a robust classification poses challenge to a rigorous estimation of the entropies .
recently , a k - mean clustering method is proposed to deliver an optimized discrete k - state classification model@xcite . in this method , the distribution of torsional angles are naturally classified in to @xmath1 clusters with irregular boundaries . in this way
, it achieves an optimized classification and a high silhouette value , indicating very good quality of clustering , is obtained . in this paper , we propose a multiscale persistent function for biomolecular structure characterization and conformation entropy evaluation .
our multiscale persistent function is based on two major methods , i.e. , persistent homology and flexibility rigidity index ( fri ) .
persistent homology is deeply rooted in algebraic topology but finds great potential in the simplification of complex data @xcite .
unlike the traditional topological method , persistent homology provides a multiscale topological representation .
it is able to embed a geometric measurement into topological invariants and provides a bridge between geometry and topology .
the essence of persistent homology is filtration . by the consistent variation of a certain filtration parameter ,
a series of topological spaces from various scales are generated . during the filtration , topological invariants , i.e.
, betti numbers , will be generated and further continue or persist for some time .
this life length or persistent time gives a relative geometric measurement of the associated topological properties .
persistent homology finds great success in topological characterization identification and analysis ( cia ) and has been used in a variety of fields , including shape recognition @xcite , network structure @xcite , image analysis @xcite , data analysis @xcite , chaotic dynamics verification @xcite , computer vision @xcite and computational biology @xcite .
recently , we have introduced persistent homology for structure characterization and mathematical modeling of fullerene molecules , proteins and other biomolecules @xcite .
consistent barcode patterns are extracted and defined as molecular topological fingerprint ( mtf ) , which is used in the analysis of protein flexibility and protein folding @xcite .
we have also developed multiresolution and multidimensional persistence @xcite .
flexibility rigidity index is originally proposed to study the flexibility and dynamic modes of biomolecules , particularly for the prediction of debye - walter factors or b - factors@xcite .
the basic assumptions of the fri method is that biomolecular structure is the equilibrium state determined by interactions within the structure .
simply speaking , if two atoms or residues are close to each other,general interaction " between them will be relatively strong .
if they are far from each other , general interaction " will be relatively week . in this way
, a distance - based monotonically - decreasing function is chosen to describe the particle - particle interactions@xcite . here
a particle can be an atom , a residue or a certain domain .
a rigidity index is defined on each particle as the summation all of interactions between this particular particle and all the others . and
flexibility index defined as the reciprocal of rigidity index is found to be linearly proportional to the experimental b - factors .
more interestingly , our flexibility index can also be viewed as the closeness centrality in graph theory .
the physical insight is that atoms located in the center will tend to have more neighbors thus more rigid , while atoms located far from the center will have less companions thus less rigid and more flexible .
a rigidity function can be constructed to simulate the biomolecular density @xcite .
the essential idea of our multiscale persistent functions is the combination of persistent homology analysis with our multiscale rigidity function , to construct a series of multiscale persistent functions , particularly multiscale persistent entropies , for structure characterization . unlike the previous persistent entropy or topological entropy@xcite
, we incorporate a resolution parameter into our rigidity function and delivers a topology based multiscale entropy model .
based on the density filtration , our method incorporates in it a classification scheme .
the classification results are reflected in @xmath2 barcodes , and are further used in persistent entropy@xcite calculation .
we have systematically compared our entropy model with the traditional model with fixed grid sizes and found that similar results can only be obtained in two extreme situations , i.e. , the grid size and resolution parameter value are very small or very large . in the middle range ,
different results are obtained as our classification is different .
our method is tested for protein structure classification .
we found that it is able to deliver a very nice classification of all - alpha , all - beta and mixed alpha and beta proteins .
the all - alpha and all - beta proteins are dramatically different and can be clearly separated by their conformation entropies .
the mixed alpha and beta proteins have both alpha - helix and beta - sheet secondary structures .
their entropies are largely distributed between the border of all - alpha and all - beta entropies .
the results are comparable with our previous ones from machine learning method @xcite .
more interestingly , based on our multiscale persistent entropy , a protein structure index ( psi ) is proposed to describe the regularity " of protein structures .
the essential idea of psi is to evaluate disorderliness in the angle distributions .
simply speaking , for a highly regular " structure element like alpha - helix , its dihedral and bond angles are very consistent thus contribute very little to the total entropy . loops and intrinsically disorder regions are extremely irregular " in terms of dihedral and bond angles and tend to contribute a large weight in the total entropy . in this way , through the entropy evaluation , psi is able to provide a new way of structure regularity characterization .
essentially , psi can be used to describe the regularity information in any systems .
the paper is organized as following , section [ sec : theory ] is devoted to method introduction .
we give a brief introduction of conformation entropy in section [ sec : conformation_entropy ] . in this section , the discrete representation of protein structure and angle - based conformation entropy evaluation
are discussed .
multiscale rigidity function is introduced in section [ sec : multiscale_rigidity_function ] .
rigidity function is essentially a continuous version of rigidity index , and our multiscale representation is achieved through tuning a resolution parameter . the persistent homology analysis ,
is reviewed in section [ sec : pha ] , which includes the basic theory of simplicial complex , filtration , complex construction and persistence . section [ sec : mph ] is devoted for the multiscale persistent homology and multiscale persistent functions .
we propose a density based filtration over the multiscale rigidity function and define various functions on the persistent barcodes .
the multiscale persistent entropy is discussed with great details .
finally , the application of our model can be found in section [ sec : application ] .
a classification of various protein structures including all - alpha , all - beta , mixed alpha and beta , are studied in section [ sec : classification ] .
section [ sec : protein_index ] is devoted for the description of protein structure index ( psi ) .
the paper ends with a conclusion .
entropy is proposed for the characterization of disorderlyness of a system@xcite .
it measures the freedom of a system to evolve into various potential configurations .
entropy is the key property to understand a wide variety of physical , chemical , and biochemical phenomena .
it plays very important roles in various biomolecular structures and interactions including protein folding , protein - protein interaction , protein - ligand binding , chromosome configuration , dna translation and transcription , etc .
for instance , the folding of a single peptide chain into a well - defined native structure is greatly facilitated by the reduction of its conformation entropy .
therefore , conformation entropy calculation is of essential importance in computational chemistry , biophysics and biochemistry .
the evaluation of the entropy necessitates a characterization of biomolecular configuration spaces . due to the limitation of cartesian grid based representation , structural parameters , particulary
dihedral angles and bond anlges , are usually employed for structure description .
their distributions can be derived from various computational methods , including molecular dynamics , monte carlo simulation , normal mode analysis , etc .
various microstates can be obtained by the discretization of the angle distribution , usually through a equal - spacing grid . and
conformation entropy can be evaluated by using the classic shannon entropy form . even though it is suggested that the relative entropy is consistent in this procedural
, a lack of a robust classification still poses challenge to a rigorous estimation of the entropie . in this section , multiscale
persistent analysis and multiscal persistent functions for biomolecular structure characterization and conformation entropy evaluation are introduced .
this methodology is founded on two major methods , i.e. , persistent homology and flexibility rigidity index .
the great details will be given in the following sections .
but before that , a brief review of the conformation entropy is given firstly . in physics ,
the second law of thermodynamics states that the total entropy of an isolated system always increases .
this law also leads to the clausius thermodynamic definition of entropy as following : @xmath3 here @xmath4 is the amount of heat absorbed by the system . for a real biomolecular system ,
this definition of entropy is too abstract and hard to implement .
normally , a direct calculation of entropy is based on the boltzmann - gibbs - planck ( bgp ) expression @xmath5 where @xmath6 is the boltzmann constant and @xmath7 is the probability distribution function . if the potential function of a biomolecule can be modeled as @xmath8 , probability function will be linearly proportional to @xmath9 , i.e , @xmath10
. however , eq.[eq : entropy_integration ] brings challenges in practical biomolecular simulation .
this is largely due to the enormous complexity of conformation space and potential function for biomolecules . to be able to evaluate the entropy
, researchers adopt a fundamental assumption that entropy can be divided into several components@xcite . in this way
, protein folding and binding entropy can be divided into protein contribution @xmath11 and solvent contribution @xmath12 .
further , protein contribution can be subdivided into conformation entropy @xmath13 , rotational entropy @xmath14 and translation entropy @xmath15 . [ cols="^ " , ] [ tab : protein_index ]
in this paper , we discuss our multiscale persistent homology analysis , particulary multiscale persistent functions .
the multiscale persistent homology analysis is based on two methods , i.e. , multiscale rigidity function and persistent homology .
the rigidity function is essential to a continuous version of rigidity index .
it incorporate in it the multiscale information by turning a resolution parameter in kernel functions .
further , multiscale barcode representation of the data can be achieved by a density filtration over these multiscale rigidity functions .
multiscale persistent functions can be defined on these barcode spaces .
we discuss a particulary function , i.e. , persistent entropy , and illustrate its applications in protein structure classification and protein structure index representation .
there are several distinguished characteristics of our multiscale persistent entropy .
firstly , we naturally divide the data into several groups .
this subdivision is based on the general topological features of the elements , i.e . , ( @xmath16 , @xmath17 ) angle points .
secondly , the classification information is embedded into our @xmath2 barcodes .
essentially , each @xmath2 is a cluster in the data and the barlength represents the size of the cluster .
it needs to be pointed out that our barcode based clustering is comparable with other methods , such as , k - mean methods@xcite , spectral graph theory@xcite , etc . in this sense , the topological entropy obtained in some special kernels and scales , can be similar to the results from the other clustering methods
. however , topological entropies defined from high dimensional @xmath18 barcodes are dramatically different from all the previous methods .
it should be noticed that topological invariants are able to describe the global structure information , and only persistent homology is able to characterize topological invariant with a geometric measurement , i.e. , persistence . in this way
, high dimensional barcode based topological entropies will be able to provide more interesting intrinsic topological information of the structure
. this will be our future research topics .
this work was supported in part by ntu sug - m4081842.110 .
the author zhiming li thanks the chinese scholarship council for the financial support no .
201506775038 .
r. baron , p. h hunenberger , and j. a. mccammon .
absolute single - molecule entropies from quasi - harmonic analysis of microsecond molecular dynamics : correction terms and convergence properties . , 5(12):31503160 , 2009 .
a. t. hagler , p.s .
stern , r. sharon , j. m. becker , and f. naider .
computer simulation of the conformational properties of oligopeptides .
comparison of theoretical methods and analysis of experimental results .
, 101(23):68426852 , 1979 .
a. korkut and w. a. hendrickson .
stereochemistry of polypeptide conformation in coarse grained analysis . in _
biomolecular forms and functions : a celebration of 50 years of the ramachandran map _ , pages 136147 .
world scientific publishing , 2013 .
y. yao , j. sun , x. h. huang , g. r. bowman , g. singh , m. lesnick , l. j. guibas , v. s. pande , and g. carlsson .
topological methods for exploring low - density states in biomolecular folding pathways . , 130:144115 , 2009 . | in this paper , we introduce multiscale persistent functions for biomolecular structure characterization .
the essential idea is to combine our multiscale rigidity functions with persistent homology analysis , so as to construct a series of multiscale persistent functions , particularly multiscale persistent entropies , for structure characterization . to clarify the fundamental idea of our method ,
the multiscale persistent entropy model is discussed in great detail .
mathematically , unlike the previous persistent entropy or topological entropy@xcite , a special resolution parameter is incorporated into our model .
various scales can be achieved by tuning its value .
physically , our multiscale persistent entropy can be used in conformation entropy evaluation .
more specifically , it is found that our method incorporates in it a natural classification scheme .
this is achieved through a density filtration of a multiscale rigidity function built from bond and/or dihedral angle distributions . to further validate our model , a systematical comparison with the traditional entropy evaluation model
is done .
it is found that our model is able to preserve the intrinsic topological features of biomolecular data much better than traditional approaches , particularly for resolutions in the mediate range .
moreover , our method can be successfully used in protein classification . for a test database with around nine hundred proteins ,
a clear separation between all - alpha and all - beta proteins can be achieved , using only the dihedral and pseudo - bond angle information .
the persistent entropy values of mixed - alpha - and - beta proteins are also found to be in the middle region with just a few cases overlapped with the other two categories . finally , a special protein structure index ( psi ) is proposed , for the first time , to describe the regularity " of protein structures .
basically , a protein structure is deemed as regular if it has a consistent , uniformed , and orderly configuration .
our psi model is tested on a database of 110 proteins , we find that structures with larger portions of loops and intrinsically disorder regions are always associated with larger psi , meaning an irregular configuration . while proteins with larger portions of secondary structures , i.e. , alpha - helix or beta - sheet , have smaller psi .
essentially , psi can be used to describe the regularity " information in any systems .
key words : conformation entropy , dihedral angle , multiscale persistent function , protein structure , persistent homology , rigidity function . |
the collective excitations of the spins in the cuo@xmath4 planes of the high temperature superconductors display a plethora of behavior @xcite , perhaps as exotic as the bulk properties themselves .
a characterization of these excitations is motivated by several considerations .
the spin excitations , which can be measured by a well understood and easily interpretable probe such as neutron scattering , provide a window on the electronic correlations in these complex materials .
indeed , cooper pairing itself is a correlation between electron spins .
a further motivation , especially in view of the unconventional pairing @xcite in these materials , is that the magnetic excitations may be involved in the pairing attraction . in the present paper
we review our recent measurements @xcite of the dynamic response over a wide frequency range i.e. up to energies of order @xmath8 .
we find that in the superconductors there is significant spectral weight above the superconducting gap energy . for both superconductors studied
, the magnetic response @xmath9 is actually stronger for frequencies just above @xmath10 than in the parent insulating antiferromagnet .
we have been careful to convert our measurements into absolute units .
this allows a systematic comparison between different systems , compositions and theories to be made .
our experiments were performed using the het and mari spectrometers at the isis pulsed spallation neutron source of the rutherford appleton laboratory .
het and mari are direct geometry chopper spectrometers .
a pulse of neutrons of approximately 1 @xmath11s duration is produced when a 800 mev pulsed proton beam hits a tantalum target .
the neutrons are monochromated using a fermi chopper at 10 m from the target and appropriately phased to the proton pulse .
scattered neutrons are detected using 800 @xmath12he detectors 4 m from the sample .
neutron scattering directly measures the imaginary part of the generalized magnetic susceptibility @xmath13 .
the scattering cross - section for an isotropic system is @xcite , @xmath14 where @xmath15=0.2905 barn sr@xmath16 @xmath17 , @xmath18 and @xmath19 are the incident and final neutron wavevectors and @xmath20 is the magnetic form factor .
absolute unit conventions were performed by measuring the elastic incoherent scattering from a vanadium standard @xcite under the same experimental conditions and by measuring the low - frequency coherent phonon scattering from the sample @xcite . throughout this paper , we label momentum transfers @xmath21 in units of @xmath16 by their reciprocal space positions * q*=@xmath22= @xmath23 . following previous practice , we use the _ orthorhombic _ nomenclature to label reciprocal space in the la@xmath1sr@xmath2cuo@xmath3 system and _ tetragonal _ nomenclature in the yba@xmath4cu@xmath5o@xmath6 system .
this means that the cuo@xmath4 planes are parallel to ( 010 ) in la@xmath1sr@xmath2cuo@xmath3 and ( 001 ) in yba@xmath4cu@xmath5o@xmath6 .
details of the samples are given elsewhere @xcite .
in the case of the la@xmath1sr@xmath2cuo@xmath3 system , data were collected with the ( 001 ) plane coincident with the principal scattering plane of the spectrometer . for yba@xmath4cu@xmath5o@xmath6 ,
the @xmath24 plane was used .
we first discuss the parent antiferromagnets la@xmath4cuo@xmath3 and yba@xmath4cu@xmath5o@xmath25 . for the purpose of this paper , we model the systems as a set of weakly coupled cuo@xmath4 layers or bilayers . in this case , the high - frequency spin excitations can be described by the heisenberg hamiltonian for a single cuo@xmath4 layer or bilayer , @xmath26 the first term in eq . [ heisenberg ] represents the nearest - neighbor coupling between cu spins in the same cuo@xmath4 plane . the second term ( not present in the single layer compound )
represents the coupling between nearest - neighbor cu spins in different layers . in the case of the single layer compound ( @xmath27 ) , conventional spin - wave theory in the classical large-@xmath28 limit yields the transverse dynamic susceptibility @xmath29 where , @xmath30^{1/2},\end{aligned}\ ] ] and @xmath31 .
we have included a `` quantum renormalization '' of the overall amplitude @xmath32 . in the conventional linear spin - wave model applicable for large @xmath28 , @xmath33 . for small @xmath28 , quantum corrections @xcite become important
( the neel state is not a good approximation to the ground state ) leading to a renormalization of the overall scale of the spin - wave dispersion and to a reduction in magnetic response with respect to the classical spin - wave theory .
the renormalization of the overall scale can be included in the exchange constant @xmath34 , where @xmath35 is the exchange constant occurring in eq .
[ heisenberg ] . the value of @xmath32 _
can _ be obtained by neutron scattering , if measurements are placed on an absolute intensity scale .
however , @xmath36 _ can not _ be measured directly from inelastic neutron scattering and must be estimated from theory . in the case of the @xmath7 square - lattice antiferromagnet , singh @xcite and igarashi @xcite
have estimated @xmath37 and @xmath32=0.51 based on a @xmath38 expansion .
the presence of the second term ( i.e. @xmath39 ) in eq .
[ heisenberg ] leads to the existence of two branches in the spin - wave dispersion which can be labeled according to whether neighboring spins in different planes rotate in the same direction ( `` acoustic '' or `` odd '' mode ) or in opposite directions ( `` optical '' or `` even '' mode ) about their time - averaged ( ordered ) directions . in the conventional linear spin - wave approximation ,
the acoustic and optic modes have the response functions per formula unit@xcite @xmath40 and @xmath41 respectively .
the dispersion relations are @xmath42{\rm ac \\ op } } ( { \bf q } ) & = & 2 z_c j_{\parallel } \left\ { 1-\gamma^2({\bf q})+j_{\perp}/j_{\parallel } \left [ 1 \pm \gamma({\bf q } ) \right ] \right\}^{1/2},\end{aligned}\ ] ] where @xmath43 $ ] and @xmath44= 3.2 is the separation of the cuo@xmath4 planes in a bilayer . the inter - planar coupling term in eq
. [ heisenberg ] leads to no additional dispersion along the @xmath45-direction , only a modulation in the amplitude of the response which nevertheless can be used to distinguish between the two modes .
the two - dimensionality of the scattering in la@xmath4cuo@xmath3 means that we are able to cut through the spin waves at several energy transfers for a single spectrometer setting .
[ lasr300 ] shows data collected for @xmath46300 mev and @xmath47 on the mari spectrometer .
panels ( b)-(f ) show constant energy cuts along the ( 1,0,0 ) direction .
a spin - wave peak is observed near @xmath48 .
twin peaks due to spin waves propagating in opposite directions are not observed due to the poor out - of - plane resolution in the ( 001 ) direction .
the broadening in the peak at higher frequencies is due to the spin - wave dispersion . a convenient way to display the data in fig .
[ lasr300 ] is as a local- or wavevector - integrated susceptibility @xmath49 , this is shown in fig .
[ lasrlocal](a ) . the solid lines in fig .
[ lasr300](b)-(f ) and fig .
[ lasrlocal ] are fits to the spin - wave model described by eq .
[ chila ] .
quantitative analysis yields an exchange constant consistent with our previous determination @xcite @xmath50 mev and @xmath51 .
this is in agreement with the calculation of igarashi @xcite and demonstrates the importance of quantum corrections in this system . as discussed above , the bilayer nature of this material results in two collective modes with additional structure to @xmath52 as compared with the single layer material .
the optic mode has a gap at the 2-d magnetic zone centers such as @xmath53 of @xmath54 .
its intensity displays an overall modulation @xmath55 which has maxima at positions @xmath56 @xmath57 .
correspondingly , the acoustic mode has an overall modulation of @xmath58 and therefore has maxima at @xmath56 @xmath59 .
we are able to probe the spin waves at various values of @xmath60 and @xmath61 by varying the incident energy .
[ ybcolow](b)-(e ) show data collected for different energy transfers at @xmath61 values corresponding to positions where the scattering is predominately acoustic .
as in la@xmath4cuo@xmath3 , a peak is observed near the magnetic zone center , for all energy transfers investigated , due to propagating spin waves . similar cuts probing @xmath61 values where the optic mode dominates ,
are shown in fig .
[ ybcolow](g)-(j ) . in this case ,
no peak is observed at the lowest energies because these energies are below the optic gap .
a more convenient way to look at these data is in the form of a local susceptibility ( see above ) .
[ ybcolocal ] shows the contributions of the two branches to the local susceptibility extracted from the data in fig .
[ ybcolow ] .
when the data are plotted in this way , we see that the optic and acoustic contributions are equal at higher frequencies . while below @xmath62 , the contribution from the optical mode is zero within the experimental error .
a detailed analysis @xcite yields a value for the optic gap of @xmath63 mev .
a similar value for the optic gap was obtained using reactor - based instrumentation @xcite . in order to determine the exchange constant @xmath64 describing the coupling of spins in different planes
, we also need to know @xmath65 because the optic gap is @xmath66 .
the in - plane exchange coupling , @xmath65 can be obtained from a measurement of the high - frequency spin waves .
[ ybcohigh ] shows data collected with the higher incident energy @xmath46600 mev .
we note that for energy transfers above 245 mev there is little variation in the intensity , suggesting that this energy is close to the zone boundary energy .
a simultaneous resolution - corrected fit of the linear spin wave model ( eqs .
( [ chiac])-([chiop ] ) ) to all our data yields values for the exchange constants of @xmath67 @xmath68 mev and @xmath69 @xmath70 mev .
further , since our measurements are in absolute units we are able to estimate the amplitude renormalization .
we find a value of @xmath71 @xmath72 , as for la@xmath4cuo@xmath3 , again the quantum renormalization is close to the value predicted by a @xmath38 expansion ( @xmath73 ) .
of particular interest is the nature of the spin fluctuations @xcite for metallic and superconducting compositions .
we have therefore studied the compounds la@xmath74sr@xmath75cuo@xmath3 ( a composition close to optimal doping ) and yba@xmath4cu@xmath5o@xmath76 ( an underdoped superconductor ) . in both cases
we find strong high - frequency spin fluctuations in the superconducting state .
[ lasr300](h)-(l ) shows data collected for la@xmath74sr@xmath75cuo@xmath3 under the same conditions as fig .
[ lasr300 ] and in the _
same units_. doping has a dramatic effect on the magnetic excitations : at low frequencies , the peaks are broader in reciprocal space and at higher frequencies there is a large suppression of the intensity . at the highest frequencies , the peaks appear to be disappearing more rapidly in the superconductor than the insulator .
this can be seen clearly when we extract and plot the local susceptibility in fig .
[ lasrlocal](b ) .
when we include low frequency data from a reactor - based experiment , it becomes clear that the local response is peaked near 22 mev .
thus doping leads to a shift of spectral weight to intermediate energies .
the low frequency dynamics of superconducting yba@xmath4cu@xmath5o@xmath6 have been the subject of considerable investigation @xcite .
the yba@xmath4cu@xmath5o@xmath76(@xmath77 k ) sample used in the current study is of high quality : it shows a sharp `` resonance peak '' @xcite at 34 mev on entering the superconducting state and has recently been shown @xcite to develop incommensurate peaks in @xmath78 for frequencies near 25 mev .
we concentrate here on the nature of the dynamics for temperatures just above the superconducting transition temperature .
[ ybco6.6low ] shows inelastic data collected for @xmath79 k on the mari spectrometer in a similar manner to that in fig .
[ ybcolow ] .
the reader should note that the energies chosen for fig .
[ ybco6.6low ] are different from those chosen for fig .
[ ybcolow ] because of the different detector positions on the het and mari spectrometers . by assuming that the scattering has a gaussian wavevector dependence for each energy , we have extracted the frequency variation of the acoustic and optical contributions to the local susceptibility @xmath80 from data such as those shown in the figure .
[ ybco6.6local ] shows the results .
comparison of figs .
[ ybco6.6low ] and [ ybco6.6local ] with figs .
[ ybcolow ] and [ ybcolocal ] shows that the magnetic response has changed dramatically between the metal and the insulator .
this is well - known at low frequencies @xcite , as in la@xmath1sr@xmath2cuo@xmath3 , there is a large broadening of the low frequency peak in wavevector .
more dramatic is the redistribution of spectral weight in energy revealed by the local susceptibility in fig .
[ ybco6.6local ] .
inspection of fig .
[ ybcolocal ] and fig .
[ ybco6.6local ] , which are in _ the same units _ and therefore can be directly compared , shows that doping causes a shift of spectral weight from high frequencies ( above 100 mev ) into the range around 50 - 100 mev .
the gap in the optic fluctuations remains and is possibly slightly reduced .
reactor - based measurements yield a similar value for the optic gap@xcite .
our measurements of @xmath80 on la@xmath4cuo@xmath3 and yba@xmath4cu@xmath5o@xmath25 are summarized in fig .
[ lasrlocal](a ) and fig .
[ ybcolocal ] .
we note that observed @xmath80 for la@xmath4cuo@xmath3 and the acoustic branch of yba@xmath4cu@xmath5o@xmath25 are of approximately equal intensity : the units of the susceptibility are expressed per formula unit in both cases , yba@xmath4cu@xmath5o@xmath6 has two cu in the cuo@xmath4 planes per formula unit .
when we compare our measurements on the quantum antiferromagnets la@xmath4cuo@xmath3 and yba@xmath4cu@xmath5o@xmath25 , we find that , in both cases , linear spin - wave theory accounts well for the dispersion and amplitude variation . however , an overall renormalization of the amplitude @xmath32 must be included to account for the observed intensity of the one - magnon response .
our observations agree with calculations of @xmath32=0.51 based on a @xmath38 expansion .
although we are able to explain the one - magnon excitation spectrum , the @xmath38 expansion also predicts additional spectral weight @xcite which has not yet been observed .
we now turn the metals la@xmath74sr@xmath75cuo@xmath81 and yba@xmath4cu@xmath5o@xmath76 .
the response of the copper spins obeys the sum rule , @xmath82 where @xmath83 is the mean squared moment on the cu sites . in both superconductors studied here ,
doping and the consequential metal - insulator transition led to a large redistribution of spectral weight and formation of a peak in @xmath80 at intermediate frequency . in the light of the sum rule , the existence of the peak in itself is not unexpected since spectral weight is lost from the bragg peak present in the insulator and the high frequency response is strongly suppressed .
what is surprising is the relatively small energy scale over which the response is distributed compared with other paramagnetic metals .
such small electronic energy scales are , of course , characteristic of the cuprate superconductors .
99 see e.g. j. m. tranquada , _ et al .
lett . * 60 * , 156 ( 1988 ) ; j. rossat - mignot , _ et al .
_ , physica * 185 - 189 * , 86 ( 1991 ) ; h. a. mook _ et al .
70 * , 3490 ( 1993 ) ; h. f. fong _ et al .
lett . * 75 * , 316 ( 1995 ) .
d. j. scalapino , phys . rep . * 250 * , 329 ( 1995 ) .
g. aeppli _ et al .
rev . lett . * 62 * 2052 ( 1989 ) ; s. m. hayden _ et al .
lett . * 67 * , 3622 ( 1991 )
. s. m. hayden _ et al .
* 76 * , 1344 ( 1996 ) .
s. m. hayden _ et al .
* 41 * , 400 ( 1996 ) ; s. m. hayden _ et al .
b * 54 * , r6905 ( 1996 ) .
_ , unpublished .
see , e.g. r. m. white , _ quantum theory of magnetism _ , springer ser .
solid - state sci .
32 ( springer , berlin 1983 ) ch 8 . throughout this paper
we use the reduced susceptibility as discussed in this reference . c. g. windsor , _ pulsed neutron scattering _
( taylor and francis , london , 1981 ) . 0 .
et al . _ ,
b * 30 * , 2377 ( 1984 ) .
r. r. p. singh , phys .
b * 39 * , 9760 ( 1989 ) .
j. igarashi , phys .
b * 46 * , 10763 ( 1992 ) ; j. phys . : condens .
matter * 4 * , 10265 ( 1992 ) .
j. m. tranquada , , phys . rev .
b * 40 * 4503 ( 1989 ) .
d. reznik _ et al .
b * 53 * , r14741 . ( 1996 ) .
k. yamada _
et al . _ , j. phys .
. jpn . * 64 * , 2742 ( 1995 ) .
et al . _ ,
* 77 * , 5425 ( 1996 ) .
_ , preprint cond - mat/9707112 p. bourges _ et al .
b , * 56 * , 11439 ( 1997 ) .
g. aeppli _ et al .
_ , _ science _ * 278 * , 1432 ( 1997 ) .
.the table summarizes experimentally - determined spin - wave parameters for la@xmath1sr@xmath2cuo@xmath3 and yba@xmath4cu@xmath5o@xmath6 . in the case of the superconductors , @xmath65 and @xmath36
are determined from the highest frequencies studied . [ cols="<,^,^,^,^",options="header " , ] | we review recent measurements of the high - frequency dynamic magnetic susceptibility in the high-@xmath0 superconducting systems la@xmath1sr@xmath2cuo@xmath3 and yba@xmath4cu@xmath5o@xmath6 .
experiments were performed using the chopper spectrometers het and mari at the isis spallation source .
we have placed our measurements on an absolute intensity scale , this allows systematic trends to be seen and comparisons with theory to be made .
we find that the insulating @xmath7 antiferromagnetic parent compounds show a dramatic renormalization of the spin wave intensity .
the effect of doping on the response is to cause broadenings in wave vector and large redistributions of spectral weight in the frequency spectrum . |
application of the algebra of biquaternions to equations of electromagnetism has been subject of an important number of research articles and books ( see , e.g. , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite and many others ) .
the aim of this work is to present some recent results in the field concerning the usage of algebraic advantages of biquaternions for analytic and numerical solution of maxwell s system for chiral media as well as for inhomogeneous media .
compared to a considerable number of publications dealing with biquaternionic reformulations of maxwell s equations for a vacuum or for a homogeneous isotropic medium , application of biquaternions to electromagnetic models corresponding to more complicated media ( a much more challenging object for studying ) was discussed in relatively few sources ( @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) .
meanwhile the possibility of representation of maxwell s system for a vacuum in the form of a single biquaternionic equation is known since 1919 @xcite , only recently it became clear how this result can be generalized for inhomogeneous media @xcite , @xcite and for chiral media @xcite .
an appropriate quaternionic or biquaternionic reformulation of a first order system of mathematical physics opens the way for applying different methods which in many aspects preserve the algebraic power of complex analysis .
for example , it is not easy to arrive at the cauchy integral formula for holomorphic functions using two - component vector formalism or even more difficult task to develop using this formalism a holomorphic function into a taylor series
. no mathematician would consider such way of presenting complex function theory helpful or appropriate .
however , this is precisely what is happening in the study of three or four - dimensional models of mathematical physics .
compare , e.g. , the stratton - chu integrals written in their standard form ( see , e.g. , @xcite ) with their biquaternionic representation @xcite , @xcite , @xcite which is in fact a convolution of a biquaternionic fundamental solution of the maxwell operator with the electromagnetic field and it is quite evident that the latter is natural and elucidating .
the meaning of the stratton - chu formulas as a cauchy integral formula for the electromagnetic field becomes transparent and no doubt in their biquaternionic form the stratton - chu formulas could be included in a moderately advanced course of electromagnetic theory which is not the usual case up to now in spite of their central role in electrical engineering applications .
the results presented in this work are essentially quaternionic in the sense that it is not clear how they could be obtained with other techniques . in the first part
( sections 2 - 5 ) we explain a numerical method for solving electromagnetic scattering problems with an unusual for three - dimensional models precision by the aid of biquaternionic fundamental solutions which to the difference of the usually utilized matrix fundamental solutions for the maxwell equations ( see , e.g. , @xcite and @xcite ) enjoy some advantageous properties .
first of all they are not matrices but vectors ( of four components ) .
second , they have a clear physical meaning of fields generated by point sources .
third , their singularity is lower than that of fundamental solutions based on a matrix approach .
the main idea of this work is to explain how our approach works and how it can be used .
we only formulate some necessary results like those about the completeness of our systems of quaternionic fundamental solutions in appropriate functional spaces referring the interested reader to some previous publications , in particular @xcite where the corresponding proofs can be found . in sections 6 - 8
we consider the time - dependent maxwell system for chiral media , rewrite it in a biquaternionic form as a single equation and then construct explicitly a corresponding green function .
section 9 is dedicated to the time - dependent maxwell equations for inhomogeneous media .
we show that these equations can also be written as a single biquaternionic equation . in a static case
the corresponding quaternionic operator factorizes the stationary schrdinger operator .
we study relationship between solutions of these important physical equations .
let @xmath0 denote the set of complex quaternions (= biquaternions ) . each element @xmath1 of @xmath0 is represented in the form @xmath2 where @xmath3 , @xmath4 is the unit and @xmath5 are the quaternionic imaginary units : @xmath6 @xmath7 we denote the imaginary unit in @xmath8 by @xmath9 as usual . by definition
@xmath9 commutes with @xmath10 , @xmath11 .
we will use the vector representation of complex quaternions , every @xmath12 is represented as follows @xmath13 , where @xmath14 is the scalar part of @xmath1 : @xmath15 , and @xmath16 is the vector part of @xmath1 : @xmath17 @xmath18 .
complex quaternions of the form @xmath19 are called purely vectorial and can be identified with vectors from @xmath20 . the operator of quaternionic conjugation we denote by @xmath21 : @xmath22 .
let us introduce the operator @xmath23 , where @xmath24 , whose action on quaternion valued functions can be represented in a vector form as follows @xmath25 that is , @xmath26 and @xmath27
. denote @xmath28 , where @xmath29 is a complex constant .
we have the following factorization of the helmholtz operator @xcite : @xmath30 using the fundamental solution of the helmholtz operator@xmath31 ( we suppose that @xmath32 ) , the fundamental solutions @xmath33 and @xmath34 for the operators @xmath35 and @xmath36 can be obtained from ( [ factor ] ) in the following way @xmath37 we have@xmath38 where @xmath39 is the dirac delta function . from ( [ fund ] )
we obtain the explicit form of @xmath33 and @xmath34:@xmath40 here @xmath41 .
note that @xmath33 and @xmath34 are full biquaternions with @xmath42 and @xmath43 .
more information on the algebra of biquaternions and related calculus can be found in @xcite .
the operators @xmath35 and @xmath36 are closely related to the maxwell equations . consider the maxwell system for a homogeneous chiral medium ( see , e.g. , @xcite ) @xmath44 and@xmath45 where @xmath46 .
some examples of numerical values of @xmath47 for physical media can be found , e.g. , in @xcite .
we notice only that when @xmath48 we obtain the maxwell system for a homogeneous , isotropic achiral medium with the wave number @xmath29 . the vectors @xmath49 and @xmath50 in ( [ m12 ] ) and ( [ m13 ] ) are complex . consider the following purely vectorial biquaternionic functions@xmath51 it is easy to verify ( see @xcite , @xcite or @xcite ) that @xmath52 and @xmath53 satisfy the following equations @xmath54 and @xmath55 where @xmath56 if @xmath48 then @xmath57 and we arrive at the quaternionic form of the maxwell equations in the achiral case ( see ( * ? ? ? * sect .
9 ) , @xcite ) but in general @xmath58 and @xmath59 are different and physically characterize the propagation of electromagnetic waves of opposing circular polarizations . obviously the vectors @xmath49 and @xmath50 are easily recovered from @xmath52 and @xmath53:@xmath60
let @xmath61 be a sufficiently smooth closed surface in @xmath62 . here
we use the term sufficiently smooth for surfaces whose smoothness allows us to introduce the corresponding sobolev space @xmath63 for a given @xmath64 . the interior domain enclosed by @xmath61 we denote by @xmath65 and the exterior by @xmath66 .
let @xmath67 and @xmath68 be two complex vectors defined on @xmath61 .
we say that @xmath67 and @xmath68 are extendable into @xmath65 if there exist such pair of vectors @xmath49 and @xmath50 defined on @xmath69 that equations ( [ m12 ] ) and ( [ m13 ] ) are satisfied in @xmath65 and on @xmath61 we have @xmath70 and @xmath71 .
the vectors @xmath67 and @xmath68 are extendable into @xmath66 if there exist such pair of vectors @xmath49 and @xmath50 defined on @xmath72 that equations ( [ m12 ] ) and ( [ m13 ] ) are satisfied in @xmath66 , the silver - mller condition @xmath73 = o(\frac{1}{\left| x\right| } ) \label{sm}%\ ] ] is fulfilled at infinity uniformly for all directions and on @xmath61 : @xmath70 and @xmath71 . with the aid of quaternionic analysis techniques these
two introduced classes of vector functions can be completely described . for achiral media
it was done in @xcite ( see also ( * ? ? ?
11 ) ) and for chiral media in @xcite .
here we recall these results without proof .
we will need the following operators@xmath74@xmath75 which are bounded in @xmath63 for all real @xmath76 .
the function @xmath77 in ( [ sa ] ) is a biquaternion valued function , @xmath78 is the quaternionic representation of the outward with respect to @xmath65 unitary normal on @xmath61 : @xmath79 and all the products in the integrand in ( [ sa ] ) are quaternionic products . from the numerous interesting properties of the operators @xmath80 , @xmath81 and @xmath82 ( see @xcite ) we will need only the following fact let complex vectors @xmath67 and @xmath68 belong to @xmath63 , @xmath83 . then 1 . in order for @xmath67 and @xmath68 to be extendable into @xmath65 the following condition is necessary and sufficient@xmath84 or which is the same @xmath85 2 . in order for @xmath67 and @xmath68 to be extendable into @xmath66
the following condition is necessary and sufficient@xmath86 or which is the same @xmath87 now we will show how two systems of quaternionic fundamental solutions suitable for the approximation of the vector functions extendable into @xmath65 or @xmath66 can be constructed .
by @xmath88 we denote a closed surface enclosed in @xmath65 and being a boundary of a bounded domain @xmath89 , and by @xmath90 we denote a closed surface enclosing @xmath69 as shown in the figure .
[ ptb ] fig1.bmp by @xmath91 we denote a set of points densely distributed on @xmath88 , and by @xmath92 a set of points densely distributed on @xmath90 . for each of these two sets
we construct a corresponding pair of systems of quaternionic fundamental solutions .
the pair of systems@xmath93 corresponds to @xmath94 and the pair of systems@xmath95 corresponds to @xmath91 .
the following theorems show us the possibility to apply the fundamental solutions ( [ syst+ ] ) for the numerical solution of interior boundary value problems for the maxwell equations ( [ m12 ] ) , ( [ m13 ] ) , and fundamental solutions ( [ syst- ] ) for the solution of exterior problems .
[ int]@xcite let two complex vectors @xmath67 and @xmath68 belong to @xmath63 , @xmath96 , be extendable into @xmath65 and both @xmath97 and @xmath98 be not eigenvalues of the dirichlet problem in @xmath65 .
then @xmath67 and @xmath68 can be approximated with an arbitrary precision ( in the norm of @xmath99 ) by right linear combinations of the form@xmath100 and@xmath101 where @xmath102 and @xmath103 are constant complex quaternions .
[ ext]@xcite let two complex vectors @xmath67 and
@xmath68 belong to @xmath63 , @xmath96 , be extendable into @xmath66 and let both @xmath97 and @xmath98 be not eigenvalues of the dirichlet problem in @xmath89 . then @xmath67 and @xmath68 can be approximated with an arbitrary precision ( in the norm of @xmath99 ) by right linear combinations of the form@xmath104 and@xmath105 where @xmath102 and @xmath103 are constant complex quaternions .
the right linear combinations in theorem [ int ] and theorem [ ext ] are in general full quaternions . in order to ensure that they will be purely vectorial additionally to a usual boundary condition for the electromagnetic field we have to add the requirement that their scalar parts be equal to zero .
we show how this can be easily achieved on some examples of numerical realization considered in the next section .
consider the exterior boundary value problem for the maxwell equations corresponding to the model of electromagnetic scattering by a perfectly conducting body with a boundary @xmath61 .
find two vectors @xmath106 and @xmath50 satisfying ( [ m12 ] ) and ( [ m13 ] ) in @xmath66 , the condition ( [ sm ] ) at infinity and the following boundary condition@xmath107 = \overrightarrow{f}(x),\qquad x\in\gamma , \label{b1}%\ ] ] where @xmath108 is a given tangential field .
we look for the solutions in the form ( [ eext ] ) and ( [ hext ] ) , applying the collocation method in order to find the coefficients @xmath102 and @xmath103 .
substitution of the vector part of ( [ eext ] ) in ( [ b1 ] ) gives us two linearly independent equations in every collocation point . in each collocation point
@xmath109 we must require also that@xmath110 and@xmath111 which gives us other two linearly independent equations . taking into account that in ( [ eext ] ) and ( [ hext ] ) we have @xmath112 unknown complex quantities we need @xmath113 collocation points .
after having solved the corresponding system of linear algebraic equations we obtain the coefficients @xmath102 and @xmath103 and consequently the approximate solution of the problem .
a good approximation of the boundary condition ( [ b1 ] ) guarantees a good approximation of the electromagnetic field in the domain @xmath66 due to the following estimate ( see @xcite)@xmath114 where @xmath115 stands for the supremum norm in any closed subset @xmath116 of @xmath66 and @xmath117 is a positive constant depending on @xmath61 and @xmath116 .
the method was tested @xcite , @xcite , @xcite using different exact solutions .
for example , let @xmath48 and consequently @xmath118 .
the vectors @xmath119 and@xmath120 where @xmath121 is constant , represent the electromagnetic field of a magnetic dipole situated at the origin ( * ? ? ?
they satisfy ( [ m12 ] ) and ( [ m13 ] ) ( for @xmath122 as well as the silver - mller conditions at infinity .
let @xmath61 be an ellipsoid described by the equalities .
@xmath123 where @xmath124 , @xmath125 .
then @xmath126 and @xmath127 give us the solution of the following boundary value problem@xmath128@xmath129 @xmath107 = \overrightarrow{f}(x),\qquad x\in\gamma\ ] ] where @xmath130{c}% c_{3}\partial_{2}\theta_{\alpha}(x)-c_{2}\partial_{3}\theta_{\alpha}(x)\\ c_{1}\partial_{3}\theta_{\alpha}(x)-c_{3}\partial_{1}\theta_{\alpha}(x)\\ c_{2}\partial_{1}\theta_{\alpha}(x)-c_{1}\partial_{2}\theta_{\alpha}(x ) \end{array } \right ) \times\overrightarrow{n}(x)\right ] .\ ] ] we give the numerical results for @xmath131 , @xmath132 and @xmath133 in ( [ ellips ] ) .
as the auxiliary surface @xmath88 containing points @xmath134 we have chosen an ellipsoid interior with respect to @xmath61 with @xmath1 , @xmath135 and @xmath136 multiplied by @xmath137 . in the following table we present the results for @xmath138 and for different values of @xmath139
the corresponding errors represent the absolute maximum difference between the exact and the approximate solutions at the points on the ellipsoid exterior with respect to @xmath61 with @xmath1 , @xmath135 and @xmath136 multiplied by @xmath140 .
[ c]||c|c|c||@xmath139 & error for @xmath49 & error for @xmath50 + 10 & 0.441e-03 & 0.332e-03 + 15 & 0.693e-05 & 0.713e-05 + 20 & 0.162e-05 & 0.186e-05 + 25 & 0.245e-06 & 0.248e-06 + 30 & 0.113e-06 & 0.171e-06 + 35 & 0.522e-07 & 0.409e-07 + a quite fast convergence of the method can be appreciated ( all numerical results were obtained on a pc pentium 4 ) .
let us notice that the approximation by linear combinations of quaternionic fundamental solutions can be applied to other classes of boundary value problems for the maxwell system like for example the impedance problem with the boundary condition@xmath107 -\xi\left [ \left [ \overrightarrow{h}(x)\times\overrightarrow{n}(x)\right ] \times\overrightarrow{n}(x)\right ] = \overrightarrow{f}(x),\qquad x\in\gamma.\ ] ] this implies some obvious changes in the matrix of coefficients of the system of linear algebraic equations corresponding to collocation points .
more results and analysis of numerical experiments were given in @xcite .
consider time - dependent maxwell s equations @xmath141
@xmath142 @xmath143 with the drude - born - fedorov constitutive relations corresponding to the chiral media ( see , e.g. , @xcite , @xcite , @xcite ) : @xmath144@xmath145 where @xmath47 is the chirality measure of the medium .
@xmath146 are real scalars assumed to be constants .
note that the charge density @xmath147 and the current density @xmath148 are related by the continuity equation @xmath149 .
incorporating the constitutive relations ( [ dbf1 ] ) , ( [ dbf2 ] ) into the system ( [ rot1])-([div ] ) we arrive at the time - dependent maxwell system for a homogeneous chiral medium @xmath150@xmath151 @xmath152 application of @xmath153 to ( [ max1 ] ) and ( [ max2 ] ) allows us to separate the equations for @xmath49 and @xmath50 and to obtain in this way the wave equations for a chiral medium@xmath154@xmath155 it should be noted that when @xmath48 , ( [ wave1 ] ) and ( [ wave2 ] ) reduce to the wave equations for non - chiral media but in general to the difference of the usual non - chiral wave equations their chiral generalizations represent equations of fourth order .
in this section following @xcite we rewrite the field equations from section [ maxw ] in a biquaternionic form .
let us introduce the following biquaternionic operator @xmath156 and consider the purely vectorial biquaternionic function @xmath157 @xcite the equation @xmath158 is equivalent to the maxwell system ( [ max1])-([max3 ] ) , the vectors @xmath49 and @xmath50 are solutions of ( [ max1])-([max3 ] ) if and only if the purely vectorial biquaternionic function @xmath159 defined by ( [ v ] ) is a solution of ( [ amax ] ) .
the scalar and the vector parts of ( [ amax ] ) have the form@xmath160@xmath161 the real part of ( [ vec ] ) coincides with ( [ max1 ] ) and the imaginary part coincides with ( [ max2 ] ) .
applying divergence to the equation ( [ vec ] ) and using the continuity equation gives us @xmath162 taking into account these two equalities we obtain from ( [ sc ] ) that the vectors @xmath49 and @xmath50 satisfy equations ( [ max3 ] ) .
it should be noted that for @xmath48 from ( [ a ] ) we obtain the biquaternionic maxwell operator for a homogeneous achiral medium for which the following equality is valid@xmath163 in the case under consideration ( @xmath164 ) we obtain a similar result .
let us denote by @xmath165 the complex conjugate operator of @xmath166:@xmath167 for simplicity we consider now a sourceless situation . in this case the equations ( [ wave1 ] ) and ( [ wave2 ] )
are homogeneous and can be represented as follows @xmath168 where @xmath169stands for @xmath49 or for @xmath50 .
here we present a procedure from @xcite which gives us a green function for the operator @xmath166 . consider the equation@xmath170 applying the fourier transform @xmath171 with respect to the time - variable @xmath172 we obtain@xmath173 where @xmath174 the last equation
can be rewritten as follows@xmath175 where @xmath176 the fundamental solution of @xmath35 is given by ( [ fund_d ] ) , so we have@xmath177 from where @xmath178 \frac{e^{i\left\vert x\right\vert \frac{\sqrt{\varepsilon\mu}\omega}{\beta\sqrt{\varepsilon\mu}\omega-1}}}% { 4\pi\left\vert x\right\vert } .\ ] ] we write it in a more convenient form @xmath179 where @xmath180 , @xmath181 , @xmath182@xmath183 in order to obtain the fundamental solution @xmath184 we should apply the inverse fourier transform to @xmath185 . among different regularizations of the resulting integral
we should choose the one leading to a fundamental solution satisfying the causality principle , that is vanishing for @xmath186 .
such an election is done by introducing a small parameter @xmath187 in the following way @xmath188 where @xmath189 .
this regularization is in agreement with the condition @xmath32 .
we have@xmath190 where @xmath191 .
expression ( [ fund1 ] ) includes two integrals of the form @xmath192 where @xmath193 .
we have @xmath194 where the change of order of integration and summation is possible because the two necessary for this conditions are fulfilled : the series is uniformly convergent on each segment and the integrals of partial sums converge uniformly with respect to @xmath195 .
denote @xmath196 for @xmath197 and @xmath198 we obtain ( see , e.g. , ( * ? ? ?
8.7 ) ) @xmath199 where @xmath200 is the heaviside function . for all other cases , that is for @xmath197 and @xmath201 and for @xmath202 and @xmath203
we have that @xmath204 and the integrand in ( [ ik ] ) has a pole at the point @xmath205 of order @xmath206 . using a result from the residue theory ( * ? ? ?
4.3 ) we obtain@xmath207 consider@xmath208 and @xmath209 for @xmath186 we have that @xmath210 is equal to the sum of residues with respect to singularities in the lower half - plane @xmath211 which is zero because the integrand is analytic there .
thus we obtain@xmath212 substitution of this result into ( [ ik ] ) gives us@xmath213 now using the series representations of the bessel functions @xmath214 and @xmath215 ( see e.g. ( * ? ? ?
* chapter 5 ) ) we obtain @xmath216 substituting these expressions in ( [ fund1 ] ) and then in ( [ f ] ) we arrive at the following expression for @xmath77 : @xmath217 finally we rewrite the obtained fundamental solution of the operator @xmath166 in explicit form:@xmath218 let us notice that @xmath77 fulfills the causality principle requirement which guarantees that its convolution with the function from the right - hand side of ( [ amax ] ) gives us the unique physically meaningful solution of the inhomogeneous maxwell system ( [ max1])-([max3 ] ) in a whole space .
consider maxwell s equations in a nonchiral inhomogeneous medium .
thus we assume that @xmath219 and @xmath220 are functions of coordinates:@xmath221 then the maxwell system has the following form@xmath222@xmath223@xmath224@xmath225 in this section following the procedure exposed in @xcite we show that this system of equations can be written in the form of a single biquaternionic equation . equations ( [ min3 ] ) and ( [ min4 ] ) can be written as follows@xmath226 and@xmath227 combining these equations with ( [ min1 ] ) and ( [ min2 ] ) we obtain the maxwell system in the form@xmath228 and@xmath229 let us make a simple observation : the scalar product of two vectors @xmath230 and @xmath231 can be written as follows@xmath232 using this fact , from ( [ min11 ] ) and ( [ min12 ] ) we obtain the pair of equations@xmath233 and@xmath234 note that @xmath235 then ( [ min21 ] ) can be rewritten in the following form@xmath236 where@xmath237 analogously , ( [ min22 ] ) takes the form@xmath238 where@xmath239 introducing the notations@xmath240 multiplying ( [ min31 ] ) by @xmath241 and ( [ min32 ] ) by @xmath242 we arrive at the equations@xmath243 and@xmath244 where as before @xmath245 is the speed of propagation of electromagnetic waves in the medium
. equations ( [ minq1 ] ) and ( [ minq2 ] ) can be rewritten in an even more elegant form .
consider the function@xmath246 let us apply to it the biquaternionic operator @xmath247 we obtain@xmath248 applying ( [ minq2 ] ) and ( [ minq1 ] ) to the real and imaginary parts of this equation gives@xmath249 note that @xmath250 hence@xmath251 let us notice that @xmath252 where @xmath253 is the intrinsic wave impedance of the medium .
denote@xmath254 then@xmath255 from ( [ vsp4111 ] ) we obtain the maxwell equations for an inhomogeneous medium in the following form@xmath256 this equation first obtained in @xcite , @xcite is completely equivalent to the maxwell system ( [ min1])-([min4 ] ) and represents maxwell s equation for inhomogeneous media in a quaternionic form .
we formulate this as the following statement .
let @xmath245 , @xmath257 and @xmath258 be defined by ( [ definitionofcandw ] ) .
then two real - valued vectors @xmath259 and @xmath260 are solutions of the system ( [ min1])-([min4 ] ) iff the purely vectorial biquaternion @xmath261 is a solution of ( [ maxmain ] ) . note that if @xmath219 and @xmath220 are constant ( a homogeneous medium ) , equation ( [ maxmain ] ) turns into a well known ( at least since the work of c. lanczos @xcite ) biquaternionic reformulation of the maxwell system in a vacuum which was rediscovered by many researchers ( e.g. , @xcite and comments in @xcite ) . equation ( [ maxmain ] ) can be considered as a generalization of the vekua equation , well known in complex analysis , that describes generalized analytic functions @xcite .
recently in @xcite using the l. bers approach @xcite another generalization of the vekua equation was considered .
most likely some of the interesting results discussed in @xcite can be obtained for ( [ maxmain ] ) also .
their physical meaning would be of great interest .
let us consider the sourceless situation , that is we are interested in the solutions for the operator @xmath264 , where the complex quaternion @xmath265 represents @xmath266 or @xmath267 and has the form @xmath268 the scalar function @xmath77 is different from zero .
note that the study of the operator @xmath269 practically reduces to that of @xmath270 , as shown in @xcite . in the case of the operator @xmath264 ( which can be called the static maxwell operator )
the situation is quite different .
the factorization ( [ factschr3 ] ) was obtained in @xcite , @xcite in a form which required a solution of an associated biquaternionic riccati equation . in @xcite
it was shown that the solution has necessarily the form @xmath276 with @xmath77 being a solution of ( [ schr3 ] ) .
[ remfromschrtofirst]as @xmath273 in ( [ factschr3 ] ) is a scalar function , the factorization of the schrdinger operator can be written in the following form@xmath277 from which it is obvious that if @xmath273 is a solution of ( [ schr3 ] ) then the vector @xmath278 is a solution of the equation@xmath279 the inverse result is given by the next statement where the following notation is used@xmath280(x , y , z)=% % tcimacro{\dint \limits_{x_{0}}^{x}}% % beginexpansion { \displaystyle\int\limits_{x_{0}}^{x } } % endexpansion g_{1}(\xi , y_{0},z_{0})d\xi+% % tcimacro{\dint \limits_{y_{0}}^{y}}% % beginexpansion { \displaystyle\int\limits_{y_{0}}^{y } } % endexpansion g_{2}(x,\zeta , z_{0})d\zeta+% % tcimacro{\dint \limits_{z_{0}}^{z}}% % beginexpansion { \displaystyle\int\limits_{z_{0}}^{z } } % endexpansion g_{3}(x , y,\eta)d\eta+c\ ] ] ( @xmath117 is an arbitrary complex constant ) .
@xcite let @xmath284 be a nonvanishing particular solution of the equation@xmath285 with @xmath286 , @xmath287 and @xmath288 being complex valued functions , @xmath289 and @xmath290 in @xmath274 .
then for any scalar function @xmath291 the following equality holds@xmath292 where @xmath293 .
@xcite let @xmath281 be a solution of equation ( [ d+mdf ] ) in a simply connected domain @xmath295 where @xmath293 and @xmath284 be a nonvanishing particular solution of ( [ maineq3 ] ) .
then @xmath296\ ] ] is a solution of ( [ maineq3 ] ) .
notice that due to the fact that in ( [ mainfact3 ] ) @xmath297 is scalar , we can rewrite the equality in the form@xmath298 now , consider the equation @xmath299 where @xmath300 is an @xmath0-valued function . equation ( [ vekuamain3 ] ) is a direct generalization of the main vekua equation considered in @xcite .
moreover , we show that it preserves some of its important properties .
@xcite let @xmath301 be a solution of ( [ vekuamain3 ] ) .
then @xmath302 is a solution of the stationary schrdinger equation @xmath303 where @xmath304 .
moreover , the function @xmath305 is a solution of the equation @xmath306 and the vector function @xmath307 is a solution of the equation @xmath308
@xcite observe that the functions @xmath309 give us a generating quartet for the equation ( [ vekuamain3 ] ) .
they are solutions of ( [ vekuamain3 ] ) and obviously any @xmath310-valued function @xmath300 can be represented in the form@xmath311 where @xmath312 are complex valued functions .
it is easy to verify that the function @xmath300 is a solution of ( [ vekuamain3 ] ) iff@xmath313 in a complete analogy with the two - dimensional case .
denote @xmath314 then ( [ vekuamain3second ] ) can be written as follows@xmath315 which is equivalent to the equation@xmath316 the results of this section remain valid in the @xmath317-dimensional situation if instead of quaternions the clifford algebra @xmath318 ( see , e.g. , @xcite , @xcite ) is considered . the operator @xmath270 is then introduced as follows @xmath319 where @xmath320 are the basic basis elements of the clifford algebra .
khmelnytskaya k v , kravchenko v v and rabinovich v s 2002 quaternionic fundamental solutions for numerical analysis of boundary value problems for the maxwell equations .
_ _ proceedings of the conference on computational and mathematical methods on science and engineering cmmse-2002 alicante , spain , 20 - 25 sept . 2002 , vol .
ii , 193 - 201 .
khmelnytskaya k v , kravchenko v v and rabinovich v s 2003 quaternionic fundamental solutions for the numerical analysis of electromagnetic scattering problems .
_ _ zeitschrift fr analysis und ihre anwendungen , v. 22 , no .
1 , 147166 . kravchenko v g , kravchenko v v and williams b d 2001 a quaternionic generalization of the riccati differential equation .
kluwer acad .
publ . , _ clifford analysis and its applications _ , ed . by f. brackx et al .
, 143 - 154 .
sprssig w 2006 initial - value and boundary value problems with hypercomplex methods . in : _ clifford analysis and applications _ , sirkka - liisa eriksson ( editor ) , tampere university of technology , research report 82 . | we give an overview of recent advances in analysis of equations of electrodynamics with the aid of biquaternionic technique .
we discuss both models with constant and variable coefficients , integral representations of solutions , a numerical method based on biquaternionic fundamental solutions for solving standard electromagnetic scattering problems , relations between different operators of mathematical physics including the schrdinger , the maxwell system , the conductivity equation and others leading to a deeper understanding of physics and mathematical properties of the equations . |
the self - assembly of macromolecular building blocks into structures of well - defined shapes and sizes is a fundamental challenge in nanoscience.@xcite part of the promise of nanoscience is the development of sophisticated supramolecular assemblies that possess multifunctional properties and behavior beyond the capabilities of simpler molecules.@xcite biology is full of examples of such super - structures . in particular ,
microtubules are biopolymers that possess features distinct from standard synthetic polymers because the monomeric building block is a complex protein called tubulin.@xcite a major feature of microtubules not available in synthetic systems is that they are the track upon which kinesin / dynein motor proteins walk .
fast depolymerization of microtubules is an essential part of their biological function and quite distinct from the typical behavior of synthetic polymers .
these special functions and properties originate from the structural arrangement of tubulin in microtubules , which produces very stiff , chiral tubules .
understanding how the interactions between tubulin monomers yield the self - assembly of microtubules and determine their properties is a fundamental open issue .
this issue is pertinent to both biology and materials science , as the special properties of microtubules have led them to be used in synthetic systems to create new materials.@xcite these studies have stimulated interests in developing supramolecular systems that act as `` artificial microtubules '' , i.e. , possess features similar or analogous to microtubules .
a far - reaching issue inherent to nanoscience is the determination of the essential features of a monomer that self - assembles into a tubular structure mimicking microtubules as well as possessing other features of microtubules .
microtubules are both a motivation and an inspiration for our work .
that is , we want to understand both the assembly of tubulin into microtubules and the more general principles of the assembly of artificial tubular structures .
the large amount of data on microtubules helps define our models and provides something to compare to our simulation results .
the monomer of microtubules is a dimer of two proteins , @xmath0 and @xmath1-tubulin .
the dimers form protofilaments via noncovalent longitudinal bonding , and typically 13 protofilaments bind laterally and form tubules with an outer diameter @xmath2 nm.@xcite variations in the number of protofilaments and the helical pitch of microtubules assembled from free tubulin _ in vitro _
are significant.@xcite the structural variations have also been observed to occur to a smaller extent for _ in vivo _
microtubules , where other molecules help control assembly .
these results suggest that the controlled formation of tubules is not simple and understanding the limitations is important for designing synthetic mimics of microtubules .
tubular structures have been observed in synthetic systems .
recently , a few cases of supramolecular systems have been developed that self - assemble into tubules.@xcite tarabout _
constructed a wedge - shaped nanoparticle from beta - sheet - forming polypeptides including an artificial peptide.@xcite the nanoparticle consists of two layers composed of beta sheets , which form a bilayer structure with the hydrophilic groups on its top and bottom that sandwich the hydrophobic groups .
the top layer is wider than the bottom layer due to larger aromatic residues ( including artificial peptides ) at the edge of the top layer .
they showed that the tubule diameter can be controlled by chemical modifications of an aromatic residue involved in the contact of nanoparticles .
et al . _ have found conditions where polypeptide - grafted comblike polymers form tubular structures.@xcite moreover , they found that gold nanoparticles with grafted poly-(l - glutamic acid ) can form tubules under certain circumstances.@xcite in these systems , the monomer is effectively a single nanoparticle and the nanoparticle - nanoparticle interactions dictate the tubule formation .
it is this class of synthetic tubular systems that we are interested in and have developed some new understanding of the role of the interactions between nanoparticles and the control of the self - assembled tubular structures .
there is also a class of small amphiphilic molecules that form tubules typically by initially forming sheets that subsequently roll into tubules.@xcite this class of tubular structures is distinct from microtubules ( e.g. , not as stiff ) and is not the focus of this work , although there is some overlap in the underlying assembly phenomena . for example , shimizu s group developed wedge - shaped amphiphiles by covalently linking hydrophilic groups of different size to the two ends of a hydrophobic spacer.@xcite they experimentally showed that they can control the inner diameter of multilayered nanotubes by varying the length of the hydrophobic spacer.@xcite the wedge geometry does promote tubular structures , but developing these systems into artificial microtubules appears unlikely .
we have previously developed a coarse - grained model monomer with a wedge shape that can self - assemble into tubular structures with the appropriate binding interactions between monomers.@xcite the monomer has lateral interactions that promote ring formation and vertical interactions that promote filament formation .
we used molecular dynamics
( md ) simulations to study the self - assembly process and obtained a diagram of the self - assembled structures as a function of lateral and vertical interaction strengths .
since our model monomers can self - assemble into tubular structures and the resulting tubules exhibit similar structural polymorphism as microtubules , we have a model that can be used to systematically examine the physical origin of structural variations and explore ways to control the structure .
furthermore , the strength of lateral and vertical bonds between tubulin monomers clearly plays important roles in determining the structure of microtubules , but their effects are hard to probe experimentally since natural evolution only leads to one particular set of interaction strengths between tubulin.@xcite directly measuring the interactions between monomers is typically not possible because experiments usually only determine the net free energy difference that is the sum of many interactions .
our model system offers an opportunity to directly calculate the energetics and the structural dynamics of tubules in more detail than previously attempted .
the focus of this work is the physical origin of the factors controlling tubule assembly . to this end
, we have added new features to our model monomer to improve the structural control of the self - assembled tubules based on new understanding of the physical nature of the assembly process .
in particular , we focus on controlling the helicity of self - assembled tubules and have added an explicit chirality to the monomer so that the desired tubule helicity can be input from the monomer chirality .
unexpectedly , our earlier simulations produced helical tubules but with achiral monomers.@xcite we show here that this tubule helicity is a consequence of an underlying rotational symmetry that can occur through a twist deformation of protofilaments .
twist of tubules also occurs with chiral monomers and enables a range of tubule helicity about the desired value to occur , i.e. , the monomer chirality . in other words
, mismatch occurs between the chirality of building blocks and the pitch of assembled tubules .
we incorporate a lock - and - key binding mechanism for the vertical interactions between monomers to more precisely constrain the tubule helicity .
our studies show that the lock - and - key binding combined with monomers with built - in chirality can be used to achieve structural control in the self - assembly of tubules , and therefore provide a guidance on the design of building blocks that will efficiently self - assemble into controlled tubular structures .
the remainder of this paper is organized as follows .
the simulation methods are briefly described in the next section . in the results and discussion section
, we first examine the energetics of the achiral system and explain the physical origin of the helical tubules that assemble from the achiral wedges .
we then include explicit chirality in the monomer design and present the assembly and energetics for the chiral systems . finally , we discuss the results of the systems that incorporate both chirality and a lock - and - key interaction for the vertical binding .
these systems provide the best control of the assembly of tubular structures .
conclusions are included in the last section .
designed for nonhelical tubules ( i.e. , pitch @xmath3 ) .
( b ) a chiral wedge monomer @xmath4 with chirality @xmath5 ( i.e. , designed for tubules with @xmath6 ) . the chirality @xmath7 determines via @xmath8 the displacement @xmath9 between the binding sites on the left and right sides of the wedge , where @xmath10 is the wedge height .
( c ) an achiral wedge monomer with a lock - and - key configuration for the vertical binding , @xmath11 ; the vertical binding sites stick out at the bottom surface and are indented at the top surface .
( d ) a chiral wedge monomer with @xmath5 and with a lock - and - key configuration for the vertical binding , @xmath12 ; the top view shows that the vertical binding sites are buried below the top surface .
( e ) a 13_3 tubule with @xmath13 protofilaments and pitch @xmath14 ( each helix is colored differently ) .
( f ) the same tubule as in ( e ) but with a top view showing 13 protofilaments ( each protofilament is colored differently ) ; for clarity the binding sites are not shown in ( f ) . , width=288 ] the wedge - shaped building blocks are shown in fig .
[ fig_buildingblock ] .
each monomer is treated as a rigid body in simulations reported here .
the core of the monomer is a wedge deformed from a cube composed of 27 particles ( gray spheres in fig .
[ fig_buildingblock ] ) with an undeformed lattice constant @xmath15 , where @xmath16 is the particle diameter .
the deformation is imposed in such a way that 13 wedges will fit to form a closed ring .
the interaction between these core sites on two monomers is modeled as a short - range repulsion using the lennard - jones 12 - 6 potential @xmath17 $ ] , where @xmath18 is the energy scale .
this potential is cut off and shifted to 0 at @xmath19 .
the short - range repulsion models the excluded volume interaction between monomers .
the temperature @xmath20 is set as @xmath21 ( @xmath22 is the boltzmann constant ) in our simulations .
we will use @xmath23 as the energy unit .
attractions between monomers occur through 8 binding sites ( colored spheres in fig .
[ fig_buildingblock ] ) placed on the lateral and vertical faces of the wedge . a binding site on the left ( top ) face of a wedge only interacts with the binding site in the same color and on the right ( bottom ) face of another wedge and vice versa .
the bonding interaction is modeled as a soft potential @xmath24 $ ] , where @xmath25 is the binding strength and @xmath26 the interaction range .
the well depth of this potential is @xmath27 at @xmath28 .
we use @xmath29 and @xmath30 ( in the unit of @xmath23 ) to designate the strength of the lateral and vertical binding interactions , respectively .
we use this potential form because it smoothly goes to zero at @xmath31 . in our previous work and this paper ,
we fix @xmath32 .
the potential form allows for variation of @xmath26 , but we leave that for future work . since @xmath33 is isotropic , a minimum of 2 binding sites on each face is needed to break the rotational symmetry when two monomers bind , which is crucial to ensure that a pair of bonded monomers have the proper orientation . in the original design ,
the lateral binding sites are placed at @xmath34 outside the lateral face and in the middle plane of the core particles , as shown in fig .
[ fig_buildingblock](a ) .
the vertical binding sites are similarly placed with respect to the top or bottom face .
such a building block has a mirror symmetry and its achiral geometry is designed for nonhelical tubules .
this monomer is designated as @xmath35 . in this work
, we extend the wedge model to treat chiral monomers .
chirality is introduced by shifting the lateral binding sites oppositely on the left and right side of the wedge , respectively .
in general , a chiral monomer will de designated as @xmath36 , where @xmath7 designates the chirality of the monomer . for a @xmath36 monomer ,
the displacement between the lateral binding sites is @xmath37 , where @xmath10 is the wedge height .
the @xmath4 monomer is shown in fig .
[ fig_buildingblock](b ) .
we treat cases where the @xmath36 monomers are designed to assemble into tubules with pitch @xmath38 . to label tubular structures ,
we follow the literature and use @xmath39 to denote tubules with @xmath40 protofilaments and pitch @xmath41 ( counted in the unit of building blocks and also called _
helix start number _ in the literature on microtubules).@xcite an illustration of @xmath41 and @xmath40 is given in figs .
[ fig_buildingblock](e ) and ( f ) .
one surprise found in our previous simulations using the achiral @xmath35 is the preferential assembly of helical tubules with pitch 1 or even 2.@xcite new results here show that tubules with pitch mismatching the monomer chirality still assemble even for chiral monomers .
we found that one factor that can help suppress the mismatch and better control the tubule pitch is to make vertical binding stronger .
since simply increasing @xmath30 leads to kinetically trapped clusters of wedges , we instead explore a lock - and - key mechanism for the vertical binding interactions , which is introduced by modifying the location of the vertical binding sites .
these sites together with the central line of core particles are displaced vertically by @xmath42 , so that the top pair of binding sites are buried below the top surface of the wedge by @xmath43 , while the bottom pair stick out of the bottom surface by @xmath44 .
the lock - and - key modification and chirality are combined to make building blocks labeled @xmath45 , where the superscript @xmath46 stands for lock - and - key . @xmath11 and @xmath12 are shown in figs . [ fig_buildingblock](c ) and [ fig_buildingblock](d ) , respectively .
all simulations were performed with the lammps simulation package .
the equations of motion were integrated using a velocity - verlet algorithm with a time step @xmath47 , where @xmath48 is the unit of time and @xmath49 the mass of one site .
the simulations studying the self - assembly of various monomers involved 1000 wedges and were run for @xmath50 to @xmath51 .
the initial state was a low - density gas of monomers uniformly distributed in the simulation box .
the temperature of the systems was kept at @xmath52 with a langevin thermostat of damping rate @xmath53 .
the first important issue to be addressed is why achiral monomers form helical tubules . to this end
, we built @xmath54 tubules from @xmath35 monomers with @xmath41 ranging from 0 to 3 .
the tubules start in a state with straight protofilaments .
the energy distribution densities per monomer , @xmath55 , calculated for each @xmath41 from md simulations after the tubules reach equilibrium are shown in fig .
[ fig_twist](a ) .
as expected , the 13_0 tubule has the lowest mean energy .
however , the energy distribution of the 13_1 tubule has a large overlap with that of the 13_0 tubule .
even the energy distribution of the 13_2 tubule overlaps with that of the 13_0 tubule .
being so close energetically explains the frequent formation of 13_1 and 13_2 tubules by the achiral @xmath35 monomers in the assembly simulations starting with free monomers ( i.e. , the initial condition is a `` gas '' phase of wedge monomers ) .
the overlap becomes negligible as @xmath41 increases to 3 , which explains the lack of 13_3 tubules in the self - assembly products .
while the energy difference per monomer at the peaks of @xmath55 for 13_0 and 13_3 is only about 0.4 @xmath56 , the difference between the initial tubules formed in the self - assembly simulations is much larger since many monomers are involved in the nucleation and the total energy difference depends on the initial tubule size .
a single turn of 13 monomers puts the total energy difference above 5 @xmath56 . in some cases ( @xmath57 )
multiple partial turns form before the tubule state occurs , and the energy difference will be consequently much larger than 5 @xmath56 . , for equilibrated @xmath54 tubules with @xmath3 ( red ) , 1 ( green ) , 2 ( blue ) , and 3 ( black ) , from left to right .
( b ) a 13_2 tubule starting with straight protofilaments evolves into a lower - energy state with twisted protofilaments . in the images
each wedge making up the tubule is shown as a single sphere .
the helices formed via the lateral binding are shown in red and blue , respectively , and one protofilament is colored yellow to emphasize the twist transformation .
both ( a ) and ( b ) are for @xmath58 and @xmath59 . , width=264 ] a structural transformation must occur for the @xmath60 or 2 tubules to have energies that overlap with @xmath3 because helical tubules ( i.e. , @xmath61 ) with straight protofilaments have the neighboring lateral attractive sites displaced resulting in higher energies . to probe the role of twist deformation , we examined the energy and geometry for the 13_2 tubule .
starting from straight protofilaments as shown in the top image of fig .
[ fig_twist](b ) , this tubule quickly transforms into a steady state with twisted protofilaments as shown in the bottom image of fig .
[ fig_twist](b ) . during the transformation ,
the energy per monomer decreases with time by about 4 @xmath23 as shown in fig .
[ fig_twist](b )
. the energy decreases as protofilaments get twisted because the twist deformation brings the lateral interaction sites into alignment .
the protofilaments in the nonhelical 13_0 tubule do remain straight as expected .
figure [ fig_angle_illustration ] illustrates the twist transformation at the monomer level with the full geometry and local rotation of wedges visualized , showing how twisted protofilaments can help lower the energy of a helical tubule made out of @xmath35 monomers . as expected for straight protofilaments in a 13_2 tubule ( fig .
[ fig_angle_illustration](a ) dotted red line ) , the lateral binding is clearly not maximized because of the displacement between the lateral attractive sites . as shown by the dotted red lines in fig .
[ fig_angle_illustration](b ) , in the twist transformation the rotation of monomers aligns the lateral sites bringing their binding strengths close to their maxima .
consequently , the total energy of a helical tubule is reduced , which stabilizes the tubule .
thus the energy of tubules with @xmath62 can be close to the preferred @xmath38 case because the twist deformation of the protofilaments substantially alters the energy by making the lateral alignment closer to the ideal configuration .
a more detailed examination of the packing of @xmath35 wedges in helical tubules reveals the limit of twist deformation . in a twisted protofilament
, the wedges must rotate about their vertical axis so that its inner surface always points to the interior of the tubule , which introduces an offset between the vertical binding sites of two consecutive stacking wedges , as shown in fig .
[ fig_angle_illustration](c ) . for @xmath35 monomers ,
this offset is very small in a @xmath63 tubule , which makes it energetically close to the 13_0 tubule .
however , as the pitch of a helical tubule gets larger , the offset increases as a result of the increasing amount of twist .
the offset is clearly visible for a 13_2 tubule as in fig .
[ fig_angle_illustration](c ) and even more dramatic for a 13_3 tubule ( see fig .
s1 in the supplement ) . for this reason , the energy distributions for the 13_2 and 13_3 tubules shift to higher and higher values , and the 13_3 distribution
is well separated from the 13_0 one .
only so much twist can occur without resulting in an expensive mismatch between the vertical binding pairs .
the result of multiple pitch values occurring for tubules in thermal equilibrium reveals a limit of treating a microtubule as an elastic tubule within the framework of continuum elasticity , where the twist of the tubule always costs energy and induces a restoring force.@xcite in our simulations , the continuum theory breaks down because of the discrete nature of the building blocks .
the monomers are not deformed in the twist deformation , and thus there is no internal elastic cost .
all the interactions are between the surfaces , and the lateral and vertical sites play distinct roles with the energetic cost occurring primarily at the vertical contacts . while the lowest energy state is the same in either treatment ( i.e. , the tubule with pitch that matches the monomer chirality has the lowest energy ) , the simulations with discrete building blocks reveal multiple states in equilibrium and the role of the surface interactions between the monomers .
the twist deformation of protofilaments can be quantified by a twist angle @xmath64 , which is the angle between the protofilaments and the central axis of the tubule ( fig .
[ fig_angle_illustration](b ) ) . for a @xmath39 tubule ,
@xmath64 depends on both @xmath40 , @xmath41 , and @xmath7 , and their relation can be easily derived through geometrical considerations based on the packing of anisotropic objects on the curved surface of a tubule , called the lattice accommodation model in the literature.@xcite in general , for a @xmath39 tubule built from @xmath36 monomers , the twist angle @xmath64 is given by @xmath65 where @xmath66 and @xmath67 are the height and width of the wedge , respectively , and @xmath68 is the designed number of protofilaments in an ideal tubule .
values of @xmath64 for tubules assembled from @xmath35 are shown in fig .
[ fig_angle_s0 ] . in this case , eq .
( [ eq_lam ] ) indicates that @xmath64 is always 0 for tubules with @xmath3 , no matter how many protofilaments the tubule contains .
the results for @xmath69 tubules in fig .
[ fig_angle_s0 ] is consistent with this prediction . for @xmath39 tubules with @xmath70
, @xmath64 depends on @xmath40 via eq .
( [ eq_lam ] ) , as confirmed by the corresponding simulation results in fig .
[ fig_angle_s0 ] .
we have also studied the effects on @xmath64 of both @xmath29 and @xmath30 . as expected , @xmath64 is generally insensitive to either @xmath29 or @xmath30 because @xmath64 is mainly determined by the geometric features ( i.e. , width , height , and chirality ) of the building blocks as expressed in eq .
( [ eq_lam ] ) .
however , one thing to notice is that results in fig .
[ fig_angle_s0 ] are obtained with tubules starting with predetermined pitch and protofilaments that are appropriately twisted according to eq .
( [ eq_lam ] ) . in this case
the pitch does not change during the simulation and the tubule only fluctuates around the steady configuration , which is close to the starting one but generally only metastable .
the situation becomes much more complicated if we use tubules starting with straight protofilaments , where the relative strength of @xmath30 and @xmath29 plays an important role in affecting the stability of the tubules .
more details are given in the supplement ( see figs .
s2-s4 ) . of tubules formed by @xmath35 monomers and with various number ( @xmath40 ) of protofilaments and pitch values : @xmath3 ( red ) , @xmath60 ( green ) , @xmath6 ( blue ) , and @xmath14 ( black ) at @xmath71 and @xmath72 .
symbols are simulation results with error bars comparable or smaller than the symbol size .
lines represent the corresponding predictions of eq .
( [ eq_lam ] ) with @xmath10 and @xmath73 .
width=264 ] the above results show that the vertical interaction is important in limiting the range of tubule helicity since it regulates the twist of protofilaments .
yet , we have found that increasing @xmath30 to increase the energy cost of twist is insufficient to control the helicity of tubules .
energy distributions of prebuilt tubules with various pitch values only show slightly less overlap even for @xmath74 .
for example , the energy distributions for the 13_0 and 13_1 tubule shown in fig .
[ fig_overlap](a ) for @xmath75 look very similar to that in fig .
[ fig_twist ] for @xmath59 . for the 13_2 tubule the overlap with the energy distribution of the 13_0 tubule
is clearly reduced , but some overlap still exists . in general , the reduction in the energy overlap at large @xmath30 is not sufficient to change the range of tubule helicity ; tubules with @xmath61 will still self - assemble for @xmath35 monomers even for very large @xmath30 .
moreover , self - assembly simulations of free monomers with large @xmath30 actually yield kinetically trapped clusters and other defected structures.@xcite thus , these results imply that to control the tubule helicity additional features ( e.g. , interactions ) will have to be added to the monomer . , for various @xmath54 tubules : @xmath3 ( red ) , 1 ( green ) , 2 ( blue ) , and 3 ( black ) .
( a ) @xmath35 at @xmath71 and @xmath75 ; ( b ) @xmath4 at @xmath71 and @xmath76 ; and ( c ) @xmath11 at @xmath71 and @xmath75 , where the 13_3 tubule is unstable and its @xmath55 not calculated .
, width=264 ] while assembly kinetics is not the focus of this paper , it plays an important role in the self - assembly of tubules and a few important points need to be made .
the self - assembly simulations start with free monomers uniformly distributed with a low density in a box .
when @xmath77 , @xmath35 monomers tend to first form rings or helical strands , which then grow into nonhelical or helical tubules , respectively.@xcite on the contrary , if @xmath78 , then @xmath35 monomers tend to form protofilaments first , which then form curved sheets ( see examples in fig .
[ fig_tubules](f ) ) that eventually close up into tubules .
this process is similar to the proposed self - assembly pathway of tubulin into microtubules .
@xcite the final structure of the tubule is strongly affected by the closure event between the two edges of a curved sheet . for @xmath35 monomers ,
if @xmath30 is only slightly stronger than @xmath29 , then the tubule can still become helical because the sheet is flexible and thermal fluctuations can introduce an offset between the two edges of the sheet when they meet and close up . if the curved sheets were stiff enough to suppress the effects of thermal fluctuations , then the closure event would be more controlled by the intrinsic chirality of the monomer . however , for the range of interaction strengths at which wedges do self - assemble into tubules , the curved sheets are typically not stiff enough to make thermal fluctuations negligible .
the flexibility of the curved sheets impacts the number of protofilaments of a tubule as well . especially at @xmath57 ,
assembled tubules tend to have protofilament number fewer than the designed value 13 .
besides the twisting fluctuation in the sheets , there are circumferential fluctuations that tend to bring the opposite sides of the sheet into contact and closure when the number of protofilaments is only 11 or 12 , even though the monomer width is chosen to allow 13 wedges to fit from purely geometric considerations . after the closure , it is virtually impossible for other wedges or protofilaments to squeeze into the lattice to make 13-protofilament tubules . in other words ,
tubules are kinetically trapped in states with protofilaments fewer than designed .
it remains an interesting open question that how the design of wedge monomers can be tweaked to address the issue of kinetic trapping and to promote the formation of tubules containing 13 protofilaments .
one obvious possibility is to squeeze the width of wedges so that sheets containing fewer than 13 protofilaments are unlikely to close .
this direction will be explored in future work .
the previous discussion on the effects of @xmath29 and @xmath30 on the tubule twist , tubule energy distributions , and assembly kinetics provides the basis to understand the self - assembly of systems starting with free monomers , i.e. , the initial state is a gas phase
. results on various systems containing 1000 monomers are shown in fig .
[ fig_tubules ] .
here we just discuss parts ( a ) and ( b ) , which involve the @xmath35 monomer . in fig .
[ fig_tubules](a ) where @xmath79 , all assembled tubules are helical with pitch 1 or 2 .
we have performed more than one simulation for these parameters and nonhelical tubules ( @xmath3 ) do form in some cases , but this result is indicative of the overall finding that helical tubules are more common than nonhelical even though the monomer has no chirality , as long as @xmath80 . when the self - assembly is induced at @xmath81 ( fig .
[ fig_tubules](b ) ) , then the pitch is 0 or 1 , closer to the monomer chirality . in both cases @xmath40 is 11 or 12 , which is a result of the closure dynamics discussed above .
we now discuss results for the self - assembly of chiral monomers .
chirality is important for better modeling microtubules , which have pitch 3 .
chiral wedges are produced by introducing an offset along the vertical direction between the two sets of lateral attractive sites on the opposite sides of the wedge .
the amount of offset sets the value of chirality .
as noted earlier , ideally the @xmath36 monomer would yield a tubule with pitch @xmath38 .
the @xmath4 monomer is shown in fig .
[ fig_buildingblock](b ) .
we first present the energy distribution of prebuilt tubules with various pitch values .
an example is shown in fig .
[ fig_overlap](b ) for the @xmath4 monomer at @xmath71 and @xmath76 .
as expected , the @xmath6 tubule now has the lowest mean energy per monomer .
the @xmath60 and 3 tubules have the next two higher average energies per monomer .
their energy distributions show large overlaps with that of the @xmath6 tubule .
even the energy distribution of the @xmath3 tubule has some overlap with the rest .
the spread of the four distributions is smaller than for the @xmath35 case in fig .
[ fig_overlap](a ) . for the @xmath4 ( generally @xmath36 ) monomer , the protofilaments in tubules with pitch @xmath82 ( generally @xmath62 )
are twisted , similar to those in the helical tubules made of @xmath35 monomers ( fig .
[ fig_twist](b ) ) .
the twist angle is found to be consistent with the prediction of eq .
( [ eq_lam ] ) .
thus , while including chirality in the monomer shifts the lowest energy state to the tubules with the preferred pitch ( as input from the monomer chirality ) , the twist deformation of protofilaments still yields overlapping energy distributions with mismatched pitch values .
images of the tubules formed during the self - assembly simulations of @xmath35 , @xmath83 , and @xmath4 at @xmath81 are shown in figs .
[ fig_tubules](b)-(d ) , respectively . as the value of @xmath7 shifts so does the pitch @xmath41 of the assembled tubules .
the results are similar between the @xmath35 and @xmath83 cases .
there are tubules with @xmath38 formed , but there also tubules with other pitch values .
surprisingly , the self - assembly of @xmath4 monomers shows that all the assembled tubules have pitch 2 , which is the same as the built - in chirality of the @xmath4 monomers .
three independent runs with different initial conditions all produce tubules only with @xmath6 .
but our expectation is that this is just a consequence of small samples .
for the three cases in fig .
[ fig_tubules](b)-(d ) , when the interactions strengths are switched to stronger lateral binding over vertical ( @xmath84 ) , a wide range of pitch values occurs ( see table 1 in the supplement ) . particularly in the strong @xmath29 limit the tubules with mismatched pitch values
seem to dominate , indicating the effects of strong lateral bonds on the nucleation of seeding structures and the assembly pathway .
more discussion is included at the end of sec .
[ sec_lk ] .
overall , the data indicate that though the pitch of tubules can be varied in a certain range by using chiral monomers , it is difficult to achieve a precise match between the pitch and chirality .
the pitch of assembled tubules usually shows a range centered on the chirality of monomers with @xmath85 .
we have found that the twist deformation makes control of the tubule pitch difficult in most cases .
while large @xmath30 would limit or even suppress twist , it generally leads to kinetically trapped structures rather than well - defined tubules .
an additional feature or modification of the vertical interaction is required in order to prevent twist . to this end
, we introduce a lock - and - key mechanism for the vertical binding . the vertical binding sites together with the central line of core sites of the wedge were shifted along the vertical direction by @xmath42 such that on the bottom face the attractive sites stick out , while on the top face the attractive sites become buried below the surrounding core sites .
in this way a simple lock - and - key configuration is created .
the vertical lock - and - key ( lk ) modification on @xmath36 monomers is labeled @xmath45 .
images of the @xmath11 and @xmath12 monomer are shown in fig .
[ fig_buildingblock](c ) and ( d ) , respectively .
the energy distributions for prebuilt tubules of @xmath11 monomers are shown in fig .
[ fig_overlap](c ) at @xmath86 .
these interaction strengths are just strong enough to induce self - assembly for this lock - and - key monomer as determined by our assembly simulations starting with free monomers . in general
, the addition of the lock - and - key requires a larger value of @xmath30 for the monomers to bind because extra energy is needed to insert the key and to compensate for the stronger repulsion between the core sites ( see fig .
s5 and relevant discussion in the supplement ) . compared with tubules of monomers without the vertical lock - and - key binding , overlaps between the energy distributions for tubules with various pitch values are reduced in the lock - and - key case . for the @xmath11 monomer , while a significant overlap still exists for tubules with @xmath60 and @xmath3 , the overlap between @xmath6 and @xmath3 is almost completed gone and tubules with @xmath14 are not even stable anymore .
thus , tubules with the lowest energy per monomer ( e.g. , tubules with @xmath3 for @xmath11 monomers ) will be more favored during the self - assembly of lock - and - key monomers .
the self - assembly of @xmath11 monomers with the same interaction strengths as for fig .
[ fig_overlap](c ) is shown in fig .
[ fig_tubules](f ) .
again , these simulations start with a gas phase of monomers . for @xmath11 ,
all the assembled tubules are nonhelical with pitch 0 , matching the chirality of @xmath11 .
some clusters are curved sheets that do not close up into tubules yet because of the exhaustion of free monomers . however , in these sheets the protofilaments are straight , and if more monomers were supplied to the system , the tubules would be expected to be nonhelical when they eventually form . two examples of the self - assembly of chiral monomers with the lock - and - key geometry are shown in fig .
[ fig_tubules](g ) and ( h ) for @xmath87 and @xmath12 , respectively .
they show that the lock - and - key mechanism produces a match between the pitch of tubules and the chirality of monomers .
this match is better than what we expect from the energy distributions , which imply that mismatched cases should occur .
we expect that simulations of larger systems that can form many tubules would produce a distribution of pitch values . nonetheless , as the comparison between figs .
[ fig_tubules](c ) and ( g ) indicates , the addition of the lock - and - key mechanism for the vertical binding substantially shifts the assembled structures toward tubules with pitch matching the monomer chirality .
results of the self - assembly simulations of @xmath45 monomers at various combinations of @xmath29 and @xmath30 are summarized in table s2 of the supplement .
this table shows that there are some ( @xmath29 , @xmath30 ) that yield only @xmath38 tubules , but also there are cases with similar ( @xmath29 , @xmath30 ) that have @xmath88 , which is what we expect based on the energy distributions .
overall , the results are indicative of an improved control of the tubule helicity and demonstrate the basic concept that strong vertical interactions are required in order to limit the amount of protofilament twist and to control the tubule pitch
. the relative importance of vertical interactions can be seen in the comparison between figs .
[ fig_tubules](e ) and ( f ) for @xmath77 and @xmath78 , respectively , for the @xmath11 monomer . for the case with stronger lateral interactions , while the chirality range is narrower and closer to the monomer chirality with the lock - and - key binding than without ( compare figs .
[ fig_tubules](e ) and ( a ) ) , the most common tubule formed by @xmath11 has pitch @xmath60 that still does not match the monomer chirality @xmath89 .
kinetics may play an important role in producing the mismatch . for stronger lateral interactions ,
tubules typically initiate by the formation of rings . the free energy barrier between forming a nonhelical ring and a helical
one is small . even if the initial binding yields a nonhelical ring , the single bond between a pair of monomers is weak and fluctuations can break the bond to allow the two ends to slip by each other and then bind together vertically forming a helical ring ( see fig . 9 in ref .
22 ) . in this manner ,
helical tubules tend to form when @xmath77 even for the @xmath11 monomer .
more generally , tubules with @xmath62 tend to form more easily for @xmath45 monomers at @xmath77 ( see table s2 in the supplement ) .
we have studied how the interactions and molecular geometry of a simple macromolecular monomer determine the assembly into tubular structures particularly with respect to the chirality of the system .
our results have important implications in the design of macromolecular building blocks that efficiently self - assemble into tubules . in order to form tubules with a given pitch value , we introduced chirality into our model monomers such that in the ideal structure the tubule pitch would match the monomer chirality .
however , we generally find that the self - assembly from free monomers into tubules with a specified pitch value does not occur
. instead a range of pitch values emerges .
we show that the twist deformation stabilizes tubules with pitch inconsistent with the chirality of monomers by better aligning neighboring monomers so that they still bind well .
these twisted tubules can have energy distributions that substantially overlap with that of the untwisted tubule with pitch equal to the monomer chirality .
however , there is a limit to the amount of twist that yields overlapping energy distributions .
besides rotating monomers about the radial axis and aligning the lateral binding sites , twist also rotates the monomers about their vertical axis , which reduces the alignment of the vertical binding sites .
thus , to prevent twist a strong vertical interaction is required that makes protofilaments stiffer and strongly raises the cost of vertical misalignment caused by the twist .
a similar trend was found recently in the packing of filament bundles , where the ground state of a bundle of stiff filaments tend to be untwisted , while filaments with low stiffness form twisted bundles.@xcite however , in our case simply increasing the interaction strength between the vertical binding sites to make protofilaments stiffer does not solve the twist issue because the self - assembly of free monomers with large interaction strengths results in kinetically trapped clusters instead of tubules . in order to overcome this hurdle , we introduced a lock - and - key mechanism into our model monomers for the vertical interactions . in this manner the control of the tubule self - assembly was substantially improved and our assembly simulations were able to achieve a good match between the tubule pitch and monomer chirality .
our results reveal the importance of the vertical interaction strength being larger than the lateral strength in microtubules.@xcite this difference is necessary to prevent twist which would yield a wide range of structures that would not be biologically functional .
the simulations also provide new insight into the _ initial _ assembly dynamics of microtubules ; most experimental work has studied the growth of microtubule ends but not the initial nucleation of microtubules . because the vertical interaction is stronger , the tubulin dimers first form protofilaments via vertical binding and the protofilaments subsequently bind laterally into curved sheets that close to form tubules .
this route generally enhances the structural control of assembled tubules and points to the importance of precise control of the sheet - closing event .
finally , we emphasize that controlling the assembly of tubules to form a specific number of protofilaments and helicity is highly nontrivial and requires many features in the monomer .
it is thus not surprising that in cells additional constraints are imposed by other molecules ( e.g. , @xmath90-tubulin ) to achieve the degree of control observed for microtubules .
this research was supported by the u.s .
department of energy , office of basic energy sciences , division of materials sciences and engineering under award kc0203010 .
sandia national laboratories is a multi - program laboratory managed and operated by sandia corporation , a wholly owned subsidiary of lockheed martin corporation , for the u.s .
department of energy s national nuclear security administration under contract de - ac04 - 94al85000 .
30ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( )
monomers with twisted protofilaments shows the rotation of wedges along a protofilament .
, width=624 ] using a 13_3 tubule of @xmath35 monomers as an example , fig .
[ fig_twist_p3 ] illustrates the limits of twist deformation of protofilaments when the helicity of a tubule does not match the the chirality of monomers .
because of the curvature of a tubule surface , a twisted protofilament has to curl around the central axis of the tubule , which in turn requires the rotation of wedge monomers along the protofilament and introduces an offset between the vertical binding sites between two stacking neighboring wedges .
if the required offset is too large , then the tubule will have a much higher energy than the one with helicity that matches the chirality of monomers , and the tubule will be energetically unfavored .
therefore , the key to suppress the rotation of wedges and thus the twist of protofilaments is to have a large @xmath30 . equation ( 1 ) of the main text describes the twist deformation of protofilaments for all tubules built from any monomers that we have studied .
one more example is included in fig .
[ fig_angle_vlks2 ] for tubules formed by @xmath12 monomers .
in general , it is expected that eq .
( 1 ) of the main text is applicable to all tubular structure made out of identical discrete building blocks . of protofilaments in tubules formed by @xmath12 monomers and with various number ( @xmath40 ) of protofilaments and pitches : @xmath6 ( red ) , @xmath60 ( green ) , and @xmath14 ( blue ) at @xmath71 and @xmath75 .
symbols are simulation results with error bars comparable or smaller than the symbol size .
lines represent the corresponding predictions of eq .
( 1 ) of the main text with @xmath10 and @xmath73 . , width=312 ] we study the effects on the tubule twist deformation of both lateral and vertical binding interactions .
one set of results is included in fig .
[ fig_angle_interaction ] , which shows that the twist angle @xmath64 is insensitive to either @xmath29 or @xmath30 .
this is not surprising , because @xmath64 is mainly determined by the geometric features of the building blocks , as expressed in eq .
( 1 ) of the main text . however , the small changes of @xmath64 with @xmath29 or @xmath30 are still noteworthy . on the one hand ,
@xmath64 slightly decreases as @xmath30 is increased while @xmath29 is fixed , indicating that the twist deformation is slightly reduced at a larger @xmath30 .
this trend can be understood on the basis that a large @xmath30 makes the protofilaments stiffer and helps reduce the offset between the vertical binding sites of two stacking wedges in a protofilament ( see fig . [ fig_twist_p3 ] ) , and a smaller offset leads to a more gently twisted protofilament and thus a smaller twist angle .
on the other hand , @xmath64 essentially remains unchanged as @xmath29 is varied at a small fixed @xmath30 , or increases slowly with an increasing @xmath29 at a large fixed @xmath30 , which indicates that the twist deformation is slightly enhanced with a large @xmath29 .
the underlying physics is that a larger @xmath29 implies a stronger adhesion between neighboring protofilaments , which favors a twisted packing of protofilaments .
the similar trends in @xmath64 vs. @xmath29 and @xmath30 were observed for all tubules that we have built with our monomers ( achiral / chiral , with / without lock - and - key vertical binding ) . vs. interaction strength : ( a ) @xmath71 and @xmath30 is varied ; ( b ) @xmath59 and @xmath29 is varied ; ( c ) @xmath91 and @xmath29 is varied .
data are for @xmath54 tubules built from @xmath35 monomers with : @xmath3 ( red / bottom ) , @xmath60 ( green / middle ) , and @xmath6 ( blue / top ) .
lines are guides to the eye . , width=624 ] results on @xmath64 in fig . 4 of the main text , fig . [ fig_angle_vlks2 ] , and fig . [ fig_angle_interaction ]
were obtained with tubules starting with protofilaments twisted according to eq .
( 1 ) of the main text . however , if the starting state has straight protofilaments , then the pitch of tubules can change , especially when @xmath57 .
an example is shown in fig . [ fig_pretubule_transition ] for a prebuilt 13_2 tubule of @xmath35 monomers . here
the starting tubule has straight protofilaments and @xmath92 .
the tubule quickly transforms into a hybrid structure of 13_0 , 13_1 , and 13_2 tubules .
however , if we ran the simulation at @xmath77 with the same starting configuration where protofilaments are straight , then the tubule stays at 13_2 , but ends up with twisted protofilaments ( see fig . 2 of the main text for the case @xmath84 ) ,
of which the twist angle is consistent with eq .
( 1 ) of the main text .
the example in fig .
[ fig_twist_p3 ] is for a tubule with @xmath14 under @xmath93 . in that case the same final state with twisted protofilaments is achieved even by tubules starting with straight protofilaments , in contrast to the case @xmath57 .
however , these tubules are still meta - stable , though when @xmath80 the depths of the local minima are increased and their meta - stability is enhanced compared with the @xmath57 case . monomers that contains straight protofilaments .
( b ) at @xmath94 , the tubule in ( a ) evolves into a state with multiple pitches ( @xmath3 , 1 , and 2 ) and the protofilaments are twisted in portions with @xmath70 .
, width=480 ] figure [ fig_vertical_potential ] helps understand why a larger @xmath30 is needed to initiate the self - assembly of lock - and - key monomers . here the total potential energy , which is the sum of the repulsions between the gray core sites on two monomers and the attractions between the colored attractive sites ( see the methods section of the main text for more details on the wedge - wedge interactions ) , is plotted as a function of the separation of wedges that bind at their vertical surfaces .
the two wedges are aligned vertically and the interaction energy is calculated as a function of separation .
the zero separation corresponds to the state in which the two pairs of vertical binding sites ( the sites with the cyan and green color ) of two monomers overlap . because of the repulsion between the gray core sites , the minimum of the potential energy occurs at a positive separation , which is around 0.1@xmath16 for the monomers without the lock - and - key configuration
however , for the lock - and - key monomers ( see fig . 1 of the main text for the geometry ) , the location of the potential minimum shifts to a larger separation ( about @xmath95 for the lock - and - key monomers studied in this paper ) .
the underlying reason is as follows .
first , the gray core sites of two lock - and - key monomers start to interact at a larger separation compared with the case without lock - and - key and the interactions are purely repulsive . then at a given separation
the total repulsion between two lock - and - key monomers is always stronger than that between two monomers without lock - and - key , while the total attraction between monomers is the same in the two cases at the same separation . as a consequence , at a given @xmath30
the depth of the potential well of two vertically bound monomers is more shallow for the lock - and - key monomers and the location of its minimum moves to a more positive separation , which is clearly seen from the comparison at @xmath76 in fig .
[ fig_vertical_potential ] . to compensate for this reduction
, a larger @xmath30 is needed for the lock - and - key monomers to achieve the same total attraction when two wedges bind vertically .
for example , the well depth of the vertical binding of two @xmath11 monomers at @xmath96 is close to that of two @xmath35 monomers at @xmath76 , as shown in fig .
[ fig_vertical_potential ] .
we also find that when the well depth is similar , the curvature of the potential energy around its minimum becomes larger after the introduction of the lock - and - key vertical binding mechanism , which indicates that the potential becomes stiffer and the fluctuations in the vertical bonds are reduced .
tables [ table_shift ] and [ table_notch ] summarize all results from assembly simulations starting with free monomers . as noted in the main text , @xmath39 stands for tubules containing @xmath40 protofilaments organized with pitch @xmath41
. tubules with @xmath38 are shown in bold , where @xmath7 is the chirality of monomers .
the mismatch between @xmath41 and @xmath7 frequently occurs for monomers without lock - and - key vertical binding mechanism , as shown in table [ table_shift ] .
however , the mismatch can be reduced or even suppressed using monomers with vertical lock - and - key binding , especially when @xmath57 as shown in table [ table_notch ] . in general , for lock - and - key monomers with @xmath97 , @xmath39 tubules with @xmath38 dominate when @xmath57 , while @xmath39 tubules with @xmath98 dominate when @xmath80 . for @xmath99 monomers ,
@xmath100 tubules are only found when @xmath57 .
for @xmath101 monomers , tubules with @xmath6 are the most frequently assembled structures , no matter @xmath57 or @xmath80 .
@xmath30 & @xmath35 & @xmath83 & @xmath4 & @xmath101 + 3.0 & 3.9 & * 11_0 , 12_0 , * 11_1 , 12_1 & * 11_1 , 12_1 , * 11_2 & * 11_2 , 12_2 * & 11_2 + 4.2 & 2.6 & 12_1 , 11_2 , 12_2 & 12_0 , * 12_1 , * 11_2 & 12_0 , 13_0 , 11_1 , 12_1 , * 12_2 , 13_2 * & 12_1 , 11_2 , 12_2 , * 12_3 * + @xmath30 & @xmath11 & @xmath87 & @xmath12 & @xmath99 + 3.0 & 6.3 & * 11_0 , 12_0 * & * 11_1 , 12_1 * & * 11_2 , 12_2 * & 11_2 , 12_2 , * 11_3 , 15_3 * + 3.0 & 6.0 & no assembly & * 11_1 , 12_1 * & * 11_2 , 12_2 * & 11_2 , 12_2 + 3.3 & 6.3 & clusters & 12_0 , * 11_1 * & * 11_2 * & 11_2 , * 13_3 * + 3.3 & 6.0 & * 11_0 , 12_0 , * 11_1 , 12_1 & 14_0 , * 11_1 , 12_1 * & 12_1 , * 11_2 , 12_2 * & 11_2 , 12_2 , * 12_3 * + 3.3 & 5.7 & no assembly & * 12_1 * & 11_1 , * 11_2 , 12_2 , 13_2 * & 11_2 , 12_2 , 13_2 , * 11_3 * + 3.6 & 6.0 & * 12_0 , * 12_1 & * 12_1 * & 13_1 , * 11_2 , 12_2 * & 13_2 , * 12_3 , 14_3 * , 13_4 + 3.6 & 5.7 & * 11_0 , 12_0 * & * 11_1 , 12_1 * & 11_1 , 12_1 , * 12_2 * & 12_1 , 11_2 , 12_2 , 13_2 + 3.6 & 5.4 & * 12_0 , * 12_1 & * 11_1 , 12_1 * & 11_1 , * 11_2 , 12_2 * & 12_2 , * 11_3 * + 3.9 & 5.1 & * 12_0 * , 12_1 & 12_0 , * 11_1 , 12_1 , 13_1 * , 11_2 & 12_1 , * 12_2 , 13_2 * & 12_1 , 12_2 , * 12_3 * + 4.4 & 4.2 & * 12_0 * , 12_1 , 13_1 & 12_0 , 14_0 , * 11_1 , 12_1 , 13_1 * & 12_0 , 13_0 , 12_1 , 13_1 , * 12_2 , * 11_3 & 11_1 , 12_1 , 11_2 , 13_2 + 4.4 & 3.9 & * 13_0 * , 13_1 & 12_0 , * 12_1 , 13_1 * & 12_0 , 11_1 , 13_1 , * 11_2 , 12_2 * & 11_0 , 12_0 , 12_1 + 4.8 & 3.6 & * 13_0 * , 12_1 & 12_0 , 13_0 , 14_0 , * 12_1 * , 12_2 & 12_0 , 13_0 , 14_0 , 12_1 , 13_1 , * 12_2 * & 11_0 , 12_0 , 11_1 , 12_1 , 13_1 + | the efficient and controlled assembly of complex structures from macromolecular building blocks is a critical open question in both biological systems and nanoscience . using molecular dynamics simulations we study the self - assembly of tubular structures from model macromolecular monomers with multiple binding sites on their surfaces [ cheng et al . , soft matter , 2012 , * 8 * , 5666 - 5678 ] . in this work
we add chirality to the model monomer and a lock - and - key interaction .
the self - assembly of free monomers into tubules yields a pitch value that often does not match the chirality of the monomer ( including achiral monomers ) .
we show that this mismatch occurs because of a twist deformation that brings the lateral interaction sites into alignment when the tubule pitch differs from the monomer chirality .
the energy cost for this deformation is small as the energy distributions substantially overlap for small differences in the pitch and chirality . in order to control the tubule pitch by preventing the twist deformation ,
the interaction between the vertical surfaces must be increased without resulting in kinetically trapped structures . for this purpose
, we employ the lock - and - key interactions and obtain good control of the self - assembled tubule pitch .
these results explain some fundamental features of microtubules .
the vertical interaction strength is larger than the lateral in microtubules because this yields a controlled assembly of tubules with the proper pitch .
we also generally find that the control of the assembly into tubules is difficult , which explains the wide range of pitch and protofilament number observed in microtubule assembly . |
the @xmath3 resonance dominates many nuclear phenomena at energies above the pion - production threshold and plays an important role in the physics of the strong interaction .
it is almost an ideal elastic @xmath4 resonance , and decays into the nucleon and pion ( @xmath5 ) with the branching fraction about @xmath6 .
the only other ( electromagnetic ) decay channel ( @xmath7 ) contributes less than @xmath8 to the total decay width @xcite .
there is a very small mass gap ( less than @xmath9 ) between the @xmath10 and the nucleon , and the @xmath3 is taken as an explicit dynamical degree of freedom in the heavy baryon chiral perturbation theory @xcite . in this article , we calculate the strong coupling constant @xmath0 with the light - cone qcd sum rules , and study the decay width @xmath11 .
the strong coupling constants of the octet baryons with the vector and pseudoscalar mesons @xmath12 and @xmath13 have been calculated with the light - cone qcd sum rules @xcite .
the light - cone qcd sum rules carry out the operator product expansion near the light - cone @xmath14 instead of the short distance @xmath15 while the nonperturbative hadronic matrix elements are parameterized by the light - cone distribution amplitudes instead of the vacuum condensates @xcite .
the nonperturbative parameters in the light - cone distribution amplitudes are calculated with the conventional qcd sum rules and the values are universal @xcite .
the article is arranged as : in section 2 , we derive the strong coupling constant @xmath0 with the light - cone qcd sum rules ; in section 3 , the numerical result and discussion ; and section 4 is reserved for conclusion .
in the following , we write down the two - point correlation function @xmath16 , @xmath17 where the baryon currents @xmath18 and @xmath19 interpolate the octet baryon @xmath20 and decuplet baryon @xmath21 respectively @xcite , the external state @xmath22 has the four momentum @xmath23 with @xmath24 . the general form of the proton current can be written as @xcite @xmath25 \gamma_5 u_c(x ) + t \left [ u_a^t(x ) c \gamma_5
d_b(x ) \right ] u_c(x ) \right\ } \
, , \nonumber\end{aligned}\ ] ] in the limit @xmath26 , we recover the ioffe current .
if we retain the additional parameter @xmath27 and choose the ideal value , the sum rule maybe improved , in this article , we choose the ioffe current for simplicity .
the correlation function @xmath28 can be decomposed as @xmath29 due to the lorentz invariance , where the @xmath30 and @xmath31 are lorentz invariant functions of @xmath20 and @xmath32 . in this article
, we choose the tensor structure @xmath33 for analysis .
the strong coupling among the @xmath10 , @xmath20 and @xmath22 can be described by the following chiral lagrangian @xcite , @xmath34 \ , .
\end{aligned}\ ] ] basing on the quark - hadron duality @xcite , we can insert a complete series of intermediate states with the same quantum numbers as the current operators @xmath18 and @xmath19 into the correlation function @xmath28 to obtain the hadronic representation . after isolating the ground state contributions from the pole terms of the baryons @xmath20 and @xmath10 ,
we get the following result is missed in @xmath35 and we obtain too small value for the strong coupling constant @xmath36 to accommodate the experimental data . ] , @xmath37 where the following definitions have been used , @xmath38 the last identity corresponds to the phenomenological lagrangian in eq.(4 ) .
the current @xmath19 couples not only to the isospin @xmath39 and spin - parity @xmath40 states , but also to the isospin @xmath39 and spin - parity @xmath41 states . for a generic @xmath42 resonance @xmath43 @xcite , @xmath44 where @xmath45 is the pole residue and
@xmath46 is the mass .
the spinor @xmath47 satisfies the usual dirac equation @xmath48 . if we take the phenomenological lagrangian , @xmath49 which corresponds to @xmath50 , the contributions from the @xmath42 states can be written as @xmath51 where the @xmath31 are lorentz invariant functions of @xmath20 and @xmath32 .
if we choose the tensor structure @xmath52 , the @xmath43 has no contaminations . in the following ,
we briefly outline the operator product expansion for the correlation function @xmath53 in perturbative qcd theory .
the calculations are performed at the large space - like momentum regions @xmath54 and @xmath55 , which correspond to the small light - cone distance @xmath14 required by the validity of the operator product expansion approach .
we write down the `` full '' propagator of a massive light quark in the presence of the quark and gluon condensates firstly @xcite , @xmath56 \nonumber\\ & & + \cdots \ , , \end{aligned}\ ] ] then contract the quark fields in the correlation function @xmath16 with the wick theorem , and obtain the following result , @xmath57\gamma_5\gamma^\alpha \langle 0|d_c(0)\bar{u}_{c'}(x ) |\pi(p)\rangle\right.\nonumber \\ & & \left .
-2\gamma_5 \gamma^\alpha \langle 0|d_c(0)\bar{u}_{b'}(x)|\pi(p)\rangle\gamma_\mu c s^t_{aa'}(-x)c \gamma_\alpha
s_{bc'}(-x)\right\}\ , .\end{aligned}\ ] ] perform the following fierz re - ordering to extract the contributions from the two - particle and three - particle @xmath22-meson light - cone distribution amplitudes respectively , @xmath58 and substitute the hadronic matrix elements ( such as the @xmath59 , @xmath60 , etc . ) with the corresponding @xmath22-meson light - cone distribution amplitudes and @xmath61 .
] , finally we obtain the spectral density at the coordinate space .
once the spectral density in the coordinate space is obtained , we can translate it into the momentum space with the @xmath62 dimensional fourier transform , @xmath63(\alpha_u,\alpha_g,1-\alpha_u-\alpha_g ) \nonumber \\ & & -\frac{4f_\pi m_\pi^2}{\pi^2 } \int_0 ^ 1 dv \int^1_0 d\alpha_g \int_0^{1-\alpha_g } d\alpha_u \int_0^{\alpha_u } d\alpha \frac{\gamma ( 1)}{(-q^2)^ { 1}}\mid_{u=\alpha_u+v\alpha_g } \nonumber\\ & & \left[v_{\parallel}+v_{\perp}+(1 - 2v)(a_{\parallel}+a_{\perp } ) \right](\alpha,\alpha_g,1-\alpha-\alpha_g ) \nonumber\\ & & + \frac{4f_\pi m_\pi^2}{\pi^2 } \int_0 ^ 1 dv ( 1-v)\int^1_0 d\alpha_g \int_0^{\alpha_g } d\beta \int_0^{\beta } d\alpha \frac{\gamma ( 1)}{(-q^2)^ { 1}}\mid_{u=1-(1-v)\alpha_g}\nonumber\\ & & \left[v_{\parallel}+v_{\perp}+(1 - 2v)(a_{\parallel}+a_{\perp } ) \right](\alpha,\beta,1-\alpha-\beta ) \
, , \end{aligned}\ ] ] where @xmath64 and @xmath65 .
the @xmath66 is a small positive quantity , after taking the double borel transform , we can take the limit @xmath67 .
there is no contribution from terms of the form @xmath68 , while there is rather large contribution from that terms in the sum rules for the strong coupling constant @xmath69 , see the article `` v. m. braun and i. e. filyanov , z. phys . * c44 * ( 1989 ) 157 '' in ref.@xcite .
if we replace the decuplet baryon current @xmath19 with the octet baryon current @xmath70 ( interpolating the neutron ) and study the strong coupling constant @xmath69 , the feynman diagrams are quite different .
our mathematica code can be used to calculate the strong coupling constant @xmath69 and produce the terms @xmath71 .
the decuplet baryon current @xmath19 and octet baryon current @xmath18 have the dirac structures @xmath72 and @xmath73 respectively , where @xmath74 stands for the @xmath75 quark fields .
the dirac structure @xmath76 corresponds to the twist-2 light - cone distribution amplitude @xmath77 .
if we replace one of the `` full '' @xmath75 quark propagators with the quark condensate @xmath78 , the terms @xmath71 in the correlation function @xmath28 have the dirac structures @xmath79 or @xmath80 , which are chiral even , because only the perturbative part of the other `` full '' @xmath75 quark propagator has contribution .
it is not unexpected that they have no contribution to the chiral odd structure @xmath52 .
the light - cone distribution amplitudes @xmath81 , @xmath82 , @xmath83 , @xmath84 , @xmath85 , @xmath86 and @xmath87 of the @xmath22 meson are presented in the appendix @xcite , the nonperturbative parameters in the light - cone distribution amplitudes are scale dependent , in this article , the energy scale is taken to be @xmath88 .
the contributions proportional to the @xmath89 can give rise to three - particle ( and four - particle ) meson distribution amplitudes with a gluon ( and quark - antiquark pair ) in addition to the two valence quarks , their corrections are usually not expected to play any significant roles , the factorizable contribution is zero and the nonfactorizable contributions from the soft hadronic matrix elements are too small to accommodate the experimental data @xcite ; the net contributions from the three - valence particle light - cone distribution amplitudes to the strong coupling constant @xmath90 are rather small , about @xmath91 @xcite . in ref.@xcite , we observe that the contributions from the three - particle ( quark - antiquark - gluon ) light - cone distribution amplitudes are less than @xmath92 for the strong coupling constants @xmath93 . in this article , the contributions from the three - particle light - cone distribution amplitudes are about @xmath94 .
the contributions from the three - particle ( quark - antiquark - gluon ) distribution amplitudes of the mesons are always of minor importance comparing with the two - particle ( quark - antiquark ) distribution amplitudes in the light - cone qcd sum rules . in our previous work , we also study the four form - factors @xmath95 , @xmath96 , @xmath97 and @xmath98 of the @xmath99 with the light - cone qcd sum rules up to twist-6 three - quark light - cone distribution amplitudes and obtain satisfactory results @xcite . in a word
, we can neglect the contributions from the valence gluons and make relatively rough estimations in the light - cone qcd sum rules . ] . in this article , we take them into account for completeness . taking double borel transform with respect to the variables @xmath100 and @xmath101 respectively ( i.e. @xmath102}{\left[u(1-u)m_\pi^2+(1-u)q_1 ^ 2+uq_2 ^ 2\right]^n } \rightarrow \frac{m^{2(2-n)}}{m_1 ^ 2 m_2 ^ 2 } e^{-\frac{u(1-u)m_\pi^2}{m^2 } } \delta(u - u_0)$ ] , @xmath103 and @xmath104 ) , then subtract the contributions from the high resonances and continuum states by introducing the threshold parameter @xmath105 ( i.e. @xmath106}\int_0^{s_0 } ds s^{n-1}e^{-\frac{s}{m^2}}$ ] ) , finally we obtain the sum rule for the strong coupling constant @xmath36 , @xmath107(\alpha_u,\alpha_g,1-\alpha_u-\alpha_g ) \nonumber \\ & & -\frac{1}{\pi^2 } m^2e_0(x)f_\pi m_\pi^2 \left[\int_0^{1-u_0 } d\alpha_g \int^{u_0}_{u_0-\alpha_g } d\alpha_u \int_0^{\alpha_u } d\alpha \right.\nonumber\\ & & \left.+\int^1_{1-u_0 } d\alpha_g \int^{1-\alpha_g}_{u_0-\alpha_g } d\alpha_u \int_0^{\alpha_u } d\alpha\right]\frac{1}{\alpha_g } \nonumber\\ & & \left[v_{\parallel}+v_{\perp}+(1 - 2\frac{u_0-\alpha_u}{\alpha_g})(a_{\parallel}+a_{\perp } ) \right](\alpha,\alpha_g,1-\alpha-\alpha_g ) \nonumber\\ & & + \frac{1}{\pi^2 } m^2e_0(x)f_\pi m_\pi^2 ( 1-u_0)\int_{1-u_0}^1 d\alpha_g\frac{1}{\alpha_g^2 } \int_0^{\alpha_g}d\beta \int_0^{1-\beta}d\alpha\nonumber\\ & & \left.\left[v_{\parallel}+v_{\perp}+(1 - 2\frac{u_0+\alpha_g-1}{\alpha_g})(a_{\parallel}+a_{\perp } ) \right](\alpha,\beta,1-\alpha-\beta)\right\ } \ , , \end{aligned}\ ] ] where @xmath108 in the following , we present an ansatz for the spectral density at the level of quark - gluon degrees of freedom @xcite .
firstly , we perform a double borel transform for the correlation function ( which is denoted as @xmath109^\alpha}$ ] symbolically ) with respect to the variables @xmath110 and @xmath111 respectively , and obtain the result , @xmath112^\alpha } \nonumber\\ & = & \frac{m^{2(2-\alpha)}}{m_{1}^{2}m_{2}^{2}}\exp\left[-\frac{u_0(1-u_0)m_\pi^2 } { m^{2 } } \right]f ( u_{0 } ) \ , , \end{aligned}\ ] ] where the @xmath113 stand for the light - cone distribution amplitudes , @xmath114 , @xmath115 , @xmath116 .
then we introduce the corresponding spectral densities @xmath117 , @xmath118f ( u_{0})\nonumber\\ & = & \int_0^{\infty}ds_1 \int_0^{\infty}ds_2 \exp\left[-\frac{s_1 } { m_1^{2}}-\frac{s_2 } { m_2^{2 } } \right ] \rho(s_1,s_2 ) \ , , \end{aligned}\ ] ] and take a replacement @xmath119 , @xmath120 , @xmath121(\sigma_1+\sigma_2)\right\ } \,.\end{aligned}\ ] ] finally we take a double borel transform with respect to the variables @xmath122 and @xmath123 respectively , the resulting qcd spectral densities read @xmath124 the threshold parameter @xmath105 is taken as @xmath125 , where the @xmath126 and @xmath127 are the threshold parameters for the channels @xmath128 and @xmath129 respectively . the quantity @xmath130 is tiny and can be safely neglected .
our approach ( i.e. performing a double borel transform and taking a replacement @xmath106}\int_0^{s_0 } ds s^{n-1}e^{-\frac{s}{m^2}}$ ] . ) is an indirect way to obtain the same results .
the input parameters are taken as @xmath131 , @xmath132 , @xmath133 , @xmath134 , @xmath135 , @xmath136 , @xmath137 , @xmath138 , @xmath139 , @xmath140 @xcite , @xmath141 , @xmath142 @xcite , @xmath143 , @xmath144 , @xmath145 and @xmath146 @xcite . in this article
, we neglect the perturbative @xmath147 corrections to the strong coupling constant @xmath36 , and take the values of the pole residues @xmath148 and @xmath149 without perturbative @xmath147 corrections for consistency . in calculation , we observe the main uncertainties come from the two parameters @xmath150 and @xmath151 in the two - particle light - cone distribution amplitudes , as the dominant contributions come from the two - particle light - cone distribution amplitudes @xmath77 and @xmath82 , the contributions from the terms involving the three - particle ( quark - antiquark - gluon ) light - cone distribution amplitudes are of minor importance , about @xmath152 of the contribution from the term @xmath153 .
the uncertainty of the parameter @xmath150 obtained in ref.@xcite is very large , in this article , we take smaller uncertainty , say @xmath154 ( i.e. @xmath155 ) , which is the typical uncertainty in the qcd sum rules .
the values of the vacuum condensates have been updated with the experimental data for the @xmath156 decays , the qcd sum rules for the baryon masses and analysis of the charmonium spectrum @xcite , in this article , we choose the standard ( or old ) values to keep in consistent with the sum rules used in determining the non - perturbative parameters in the light - cone distribution amplitudes .
the threshold parameter @xmath105 is chosen to be @xmath157 to avoid possible contamination from the contribution of the @xmath158 baryon @xmath159 in the @xmath160 scattering amplitude @xcite .
furthermore , it is large enough to take into account the contribution of the @xmath3 .
however , the interpolating current @xmath19 has nonvanishing coupling with the isospin @xmath39 and spin @xmath161 states , the contribution from the @xmath162 state @xmath163 is included in if the @xmath163 has negative parity @xcite .
we choose the tensor structure @xmath164 to avoid the contamination .
the borel parameters are chosen as @xmath165 and @xmath166 , in those regions , the value of the strong coupling constant @xmath0 is rather stable with the variation of the borel parameter @xmath167 , which are shown in fig.1 .
taking into account all the uncertainties , finally we obtain the numerical results for the strong coupling constant @xmath168 , which are shown in fig.1 , @xmath169 for the parameter @xmath170 and @xmath171 respectively . with the borel parameter @xmath167 , @xmath172 for @xmath170 and @xmath173 for @xmath171 .
, title="fig:",width=264 ] with the borel parameter @xmath167 , @xmath172 for @xmath170 and @xmath173 for @xmath171 .
, title="fig:",width=264 ] the strong coupling constant @xmath0 has the following relation with the decay width @xmath11 , @xmath174[m_\delta^2-(m_p - m_\pi)^2]}}{2m_\delta } \ , .\end{aligned}\ ] ] if we take the experimental data as input parameter , @xmath175 @xcite , we can obtain the value @xmath176 , our numerical result @xmath177 is rather good . in the region
@xmath178 , @xmath179 @xcite .
if the radiative @xmath2 corrections to the leading perturbative terms are companied with large numerical factors , just like in the case of the qcd sum rules for the mass of the proton @xcite , @xmath180 the contributions of the order @xmath2 are large , neglecting them can impair the predictive ability .
furthermore , the pole residues @xmath181 and @xmath182 also receive contributions from the perturbative @xmath2 corrections , if we take them into account properly , we can improve the value of the strong coupling constant @xmath36 .
in this article , we calculate the strong coupling constant @xmath0 and study the strong decay @xmath183 with the light - cone qcd sum rules .
the numerical value of the strong coupling constant @xmath0 is consistent with the experimental data .
the small discrepancy maybe due to failure to take into account the perturbative @xmath2 corrections .
the light - cone distribution amplitudes of the @xmath22 meson are defined by @xcite @xmath184 where @xmath185 , @xmath186 and @xmath187 .
the light - cone distribution amplitudes of the @xmath22 meson are parameterized as @xcite @xmath188c_2^{\frac{3}{2}}(\xi)\right\}\ , , \nonumber \\
\phi_{3\pi}(\alpha_i ) & = & 360 \alpha_u \alpha_d \alpha_g^2 \left \{1 + \lambda_3(\alpha_u-\alpha_d)+ \omega_3 \frac{1}{2 } ( 7 \alpha_g - 3 ) \right\ } \ , , \nonumber\\ v_{\parallel}(\alpha_i ) & = & 120\alpha_u \alpha_d \alpha_g \left ( v_{00}+v_{10}(3\alpha_g-1)\right)\ , , \nonumber \\
a_{\parallel}(\alpha_i ) & = & 120 \alpha_u \alpha_d \alpha_g a_{10 } ( \alpha_d-\alpha_u)\ , , \nonumber\\ v_{\perp}(\alpha_i ) & = & -30\alpha_g^2 \left\{h_{00}(1-\alpha_g)+h_{01}\left[\alpha_g(1-\alpha_g)-6\alpha_u \alpha_d\right ] \right .
\nonumber\\ & & \left .
+ h_{10}\left [ \alpha_g(1-\alpha_g)-\frac{3}{2}\left(\alpha_u^2+\alpha_d^2\right)\right]\right\}\ , , \nonumber\\ a_{\perp}(\alpha_i ) & = & 30 \alpha_g^2 ( \alpha_u-\alpha_d ) \left\{h_{00}+h_{01}\alpha_g+\frac{1}{2}h_{10}(5\alpha_g-3 ) \right\ } , \nonumber\\ a(u)&=&6u(1-u)\left\ { \frac{16}{15}+\frac{24}{35}a_2 + 20\eta_3+\frac{20}{9}\eta_4 \right .
\nonumber \\ & & + \left [ -\frac{1}{15}+\frac{1}{16}-\frac{7}{27}\eta_3\omega_3-\frac{10}{27}\eta_4\right]c^{\frac{3}{2}}_2(\xi ) \nonumber\\ & & \left.+\left [ -\frac{11}{210}a_2-\frac{4}{135}\eta_3\omega_3\right]c^{\frac{3}{2}}_4(\xi)\right\}+\left\ { -\frac{18}{5}a_2 + 21\eta_4\omega_4\right\ } \nonumber\\ & & \left\{2u^3(10 - 15u+6u^2 ) \log u+2\bar{u}^3(10 - 15\bar{u}+6\bar{u}^2 ) \log \bar{u } \right . \nonumber\\ & & \left
. + u\bar{u}(2 + 13u\bar{u})\right\ } \
, , \nonumber\\ g(u)&=&1+g_2c^{\frac{1}{2}}_2(\xi)+g_4c^{\frac{1}{2}}_4(\xi)\ , , \nonumber\\ b(u)&=&g(u)-\phi_\pi(u)\ , , \end{aligned}\ ] ] where @xmath189 @xmath190 , and @xmath191 , @xmath192 , @xmath193 , @xmath194 are gegenbauer polynomials , @xmath195 and @xmath196 @xcite .
this work is supported by national natural science foundation , grant number 10405009 , 10775051 , and program for new century excellent talents in university , grant number ncet-07 - 0282 .
the author would like to thank prof.t.huang and prof.v.m.braun for valuable discussions and comments .
t. m. aliev , a. ozpineci and m. savci , phys .
* d64 * ( 2001 ) 034001 ; t. doi , y. kondo and m. oka , phys . rept .
* 398 * ( 2004 ) 253 ; t. m. aliev , a. ozpineci , s. b. yakovlev and v. zamiralov , phys . rev .
* d74 * ( 2006 ) 116001 ; z. g. wang , phys . rev .
* d75 * ( 2007 ) 054020 ; and references therein .
i. i. balitsky , v. m. braun and a. v. kolesnichenko , nucl . phys .
* b312 * ( 1989 ) 509 ; v. l. chernyak and i. r. zhitnitsky , nucl . phys . *
b345 * ( 1990 ) 137 ; v. l. chernyak and a. r. zhitnitsky , phys .
* 112 * ( 1984 ) 173 ; v. m. braun and i. e. filyanov , z. phys .
* c44 * ( 1989 ) 157 ; v. m. braun and i. e. filyanov , z. phys .
* c48 * ( 1990 ) 239 .
m. a. shifman , a. i. vainshtein and v. i. zakharov , nucl . phys .
* b147 * ( 1979 ) 385 , 448 ; l. j. reinders , h. rubinstein and s. yazaki , phys
. rept . * 127 * ( 1985 ) 1 ; s. narison , qcd spectral sum rules , world scientific lecture notes in physics * 26 * ( 1989 ) 1 .
b. l. ioffe , nucl .
b188 * ( 1981 ) 317 ; b. l. ioffe and a. v. smilga , nucl . phys . *
b232 * ( 1984 ) 109 ; v. m. belyaev and b. l. ioffe , sov .
jetp * 56 * ( 1982 ) 493 ; v. m. belyaev and b. l. ioffe , sov .
jetp * 57 * ( 1983 ) 716 .
v. a. beilin and a. v. radyushkin , nucl . phys . *
b260 * ( 1985 ) 61 ; p. ball and v. m. braun , phys . rev .
* d49 * ( 1994 ) 2472 ; v. a. nesterenko and a. v. radyushkin , phys . lett . *
b115 * ( 1982 ) 410 ; h. kim , s. h. lee and m. oka , prog .
* 109 * ( 2003 ) 371 .
z. h. li , t. huang , j. z. sun and z. h. dai , phys . rev .
* d65 * ( 2002 ) 076005 .
b. l. ioffe and k. n. zyablyuk , nucl . phys .
* a687 * ( 2001 ) 437 ; b. v. geshkenbein , b. l. ioffe and k. n. zyablyuk , phys . rev .
* d64 * ( 2001 ) 093009 ; b. l. ioffe and k. n. zyablyuk , eur .
j. * c27 * ( 2003 ) 229 ; k. zyablyuk , jhep * 0301 * ( 2003 ) 081 . | in this article , we calculate the strong coupling constant @xmath0 and study the strong decay @xmath1 with the light - cone qcd sum rules .
the numerical value of the strong coupling constant @xmath0 is consistent with the experimental data .
the small discrepancy maybe due to failure to take into account the perturbative @xmath2 corrections .
+ zhi - gang wang + department of physics , north china electric power university , baoding 071003 , p. r. china pacs numbers : 13.30.-a ; 13.75.gx |
quantum computation @xcite has made an enormous leap from a far - fetched promise @xcite to a realistic near - future technology @xcite during the past three decades . among others , spin systems in the solid state @xcite
have been developed into a mature but still very fast - evolving research field . in recent years increased research efforts have focused on c , si , and ge @xcite , which can be purified to only consist of isotopes with zero nuclear spin @xcite and thus exhibit exceptionally long spin lifetimes @xcite .
[ [ section ] ] the one - dimensional character of electrostatically defined quantum dots in ge - si core - shell nanowires leads to unique electronic properties in the valence band , where heavy and light hole states are mixed @xcite .
the band mixing gives rise to an enhanced rashba - type spin - orbit interaction @xcite , leading to strongly anisotropic and electric - field dependent @xmath0-factors @xcite .
this makes quantum dots in ge - si core - shell nanowires promising candidates for robust spin - orbit qubits that can be electrically controlled via circuit quantum electrodynamics @xcite .
[ [ section-1 ] ] despite these profound theoretical contributions , only few experiments in ge - si core - shell nanowires have been reported including josephson junctions @xcite , spin - filling @xcite , spin relaxation @xcite , spin coherence @xcite , charge sensing @xcite in the many - hole regime , and signatures of weak antilocalization @xcite .
[ [ section-2 ] ] in this work we experimentally explore the an - iso - tro - py of the @xmath0-factor in ge - si core - shell nanowires .
we electrostatically define a highly - tunable , elongated hole quantum dot in the nanowire by means of local gates .
we measure the zeeman splitting of a single - particle state in the quantum dot while rotating the magnetic field around the high - symmetry axes of the system and find a strong anisotropy with respect to the nanowire as well as to the electric field , in line with theoretical predictions @xcite .
our device in fig .
[ fig1](a ) consists of a p@xmath2-doped si substrate covered with 200 nm sio@xmath3 , on which six bottom gates with 100 nm pitch are patterned with electron - beam lithography ( ebl ) .
the gates are buried by 10 nm al@xmath3o@xmath4 grown with atomic layer deposition at @xmath5c . a single nanowire with a si shell thickness of @xmath62.5 nm and a defect - free ge core with a radius of @xmath68 nm @xcite is deterministically placed on top of the gate structure with a micromanipulator and then contacted with ohmic contacts made of 0.5/50 nm ti / pd .
a source - drain bias voltage @xmath7 is applied to the source , the current @xmath8 is measured between the drain and ground .
all measurements are performed using dc electronic equipment in a dilution refrigerator with a base temperature of 8 mk and an effective hole temperature of @xmath9 mk determined by measuring the temperature dependence of the coulomb peak width @xcite .
( dark red ) , six bottom gates g1-g6 ( light red ) , each @xmath635 nm wide and with 100 nm pitch buried under 10 nm al@xmath3o@xmath4 ( yellow ) , on top of which the nanowire is placed ( green ) with ohmic contacts ( 0.5/50 nm ti / pd , blue ) .
( c ) current @xmath8 vs. @xmath10 with g3 and g5 forming tunnel barriers ( @xmath11 mv , @xmath12 mv ) .
black curve is taken at @xmath13 mv , red curve at @xmath14 mv .
( d ) numerical differential conductance @xmath15 plotted vs. @xmath7 and @xmath10 at the same barrier voltages as in ( c).[fig1 ] ] [ [ section-3 ] ] we use this gate design @xcite to electrostatically define a single quantum dot @xcite .
the two barrier gates g3 and g5 control the tunnel barriers , and a third plunger gate g4 the electrochemical potential of the quantum dot . in fig . [ fig1](c ) we plot @xmath8 versus the voltage on g4 @xmath10 .
when applying a high @xmath14 mv we observe a strong suppression of @xmath8 for @xmath16 v , indicating depletion of the nanowire at @xmath17 v. at low @xmath13 mv we observe coulomb peaks @xcite with a regular spacing of @xmath18 mv over a range of 2 v , i.e. we are able to change the hole occupation of the quantum dot by more than 60 holes . above @xmath19 v no regular coulomb peaks are observed , but the high - bias gate sweep suggests that the quantum dot is not completely emptied , i.e. in this device we are unable to identify the last hole on the quantum dot . if we assume the plunger gate coupling to stay constant and the quantum dot to be empty at @xmath20 v , we can estimate the number of remaining holes to be @xmath21 at @xmath22 v. reaching the single - hole regime was not possible in our device .
[ [ section-4 ] ] a non - linear transport measurement is displayed in fig .
[ fig1](d ) . in this bias spectroscopy
we plot the numerical differential conductance @xmath23 vs. @xmath7 and @xmath10 , as will be in all the following bias spectroscopy plots .
formation of a single quantum dot is indicated by regularly shaped , closing coulomb diamonds @xcite .
the height of the coulomb diamonds indicates an addition energy of @xmath24 mev .
the variations in @xmath25 can not be explained by an interacting second quantum dot , which would lead to non - closing diamonds .
orbital shell filling can cause the variations in @xmath25 @xcite .
the low number of residing holes ( @xmath6 25 - 30 ) supports this reasoning .
[ [ section-5 ] ] the results in fig .
[ fig1 ] show a highly tunable nanowire device in which we intentionally define a very stable quantum dot .
we can control the number of holes in the quantum dot over a wide range from approximately 85 down to approximately 17 .
color scale applies to all bias spectroscopies in this figure .
( b ) @xmath26 vs. @xmath7 and @xmath27 along the green line in ( a ) .
( c ) @xmath28 extracted from b ) vs. @xmath27 together with a linear fit ( red line ) that yields @xmath29 .
( d ) bias spectroscopies measured at @xmath30 ( upper panel ) , and @xmath31 t ( lower panel ) .
the green arrows indicate the spin - degenerate and spin - split orbital ground state .
measurements in fig .
[ fig3 ] were taken along the blue line.[fig2 ] ] we now investigate the zeeman splitting @xmath28 of the spin - degenerate quantum dot states @xcite . to determine the @xmath0-factor with high accuracy , we choose a charge transition where the onset of conductance is sufficiently separated from other lines of increased conductance corresponding to , e.g. , ( orbital ) excited states of the quantum dot , or resonances due to the low dimensionality of the leads @xcite . in fig .
2(a ) we show a bias spectroscopy of such a charge transition together with the two coulomb diamonds adjacent to it .
the number of residing holes here is approximately 35 , again estimated by comparing the current plunger gate voltage ( @xmath32 mv ) to the depletion voltage at high bias ( @xmath33 v ) .
[ [ section-6 ] ] we keep the plunger - gate voltage at @xmath34 mv and sweep @xmath7 along the green line in fig . [ fig2](a ) at different magnetic fields @xmath27 while measuring the current [ fig .
[ fig2](b ) ] .
the magnetic field here is applied in the plane of the chip perpendicular to the nanowire axis . at @xmath30 , one
very pronounced peak marks the onset of conductance , which splits up into a spin - ground and spin - excited state at finite magnetic fields .
note that the shifts of the two states are symmetric and linear up to at least 2 t , indicating that for magnetic fields @xmath35 t the linear zeeman splitting is the only relevant term , and other effects , such as a diamagnetic shift @xcite , are negligible .
the spin splitting of the orbital ground state is further confirmed by the two bias spectroscopies in fig .
[ fig2](d ) at @xmath30 and @xmath31 t. the spin - degenerate orbital ground state of the charge transition at @xmath30 [ indicated by a single green arrow fig . [ fig2](d ) ] is clearly split into two lines at @xmath31 t ( indicated by two green arrows ) .
[ [ section-7 ] ] we extract the zeeman splitting @xmath28 by converting the @xmath7 scale into energy .
the lever arm @xmath36 with @xmath37 , where @xmath38 is the total capacitance of the dot , and @xmath39 is the source capacitance ) for this conversion is graphically extracted : the slopes of the coulomb diamond edges from fig .
[ fig2](a ) are @xmath40 and @xmath41 , where @xmath42 is the gate capacitance @xcite .
by using @xmath43 we find a lever arm of @xmath44 .
the linear increase of @xmath28 with increasing @xmath27 is shown in fig .
[ fig2](c ) .
we fit the slope of @xmath28 according to @xmath45 , where @xmath1 is the effective @xmath0-factor , and @xmath46 is the bohr magneton [ see fig .
[ fig2](c ) ] .
this yields an effective @xmath0-factor for this transition of @xmath29 .
we point out that the spin states are mixtures of heavy and light hole states and therefore @xmath47 , which is accounted for by the introduction of @xmath1 as an effective @xmath0-factor .
note that @xmath1 may differ significantly from transition to transition due to the varying heavy - hole light - hole mixing of subbands and quantum dot states @xcite at the valence band edge of the nanowire @xcite .
[ [ section-8 ] ] in summary , in fig .
[ fig2 ] we determine the effective @xmath0-factor @xmath1 to be @xmath29 for an in - plane magnetic field perpendicular to the nanowire .
the corresponding zeeman splitting is symmetric and linear up to at least 2 t.
to investigate the anisotropy of the @xmath0-factor , we measure the zeeman splitting of the ground state at a fixed magnetic field magnitude of @xmath48 t while changing the direction of @xmath49 .
we choose the coordinate system in accordance with maier _
@xcite , i.e. the @xmath50-axis points along the nanowire axis , the @xmath51-axis points out of the chip plane parallel to the electric field produced by the bottom gates , and the @xmath52-axis is in plane with the chip and perpendicular to the nanowire [ see fig 3(a ) ] .
we will show measurements in three orthogonal rotation planes . within each plane ,
a full @xmath53-rotation of @xmath49 is performed in steps of @xmath54 . for each step
, @xmath8 is measured vs. @xmath7 along the blue line in fig .
[ fig2](d ) .
the values for @xmath28 along the different directions are obtained by fitting the line cuts with two peaks for the spin - ground and spin - excited states and calculating the distance between the two peak centers .
[ [ section-9 ] ] first we rotate the magnetic field from the @xmath52-axis to the @xmath50-axis [ fig .
[ fig3](a ) ] . at @xmath55 ,
the zeeman splitting of @xmath56 @xmath57ev corresponds to a @xmath0-factor of @xmath58 [ in agreement with fig .
[ fig2](c ) ] .
the zeeman splitting decreases when the magnetic field is rotated towards the nanowire axis , until it is almost completely quenched at @xmath59 with @xmath60 @xmath57ev , corresponding to @xmath61 . for the magnetic field along the @xmath50-axis @xmath62
the peak is approximately twice as high and also significantly broadened compared to the @xmath63 peak .
this indicates that here the zeeman splitting is too small for the two peaks of the spin - excited and spin - ground state to be resolved .
however , the broadened peak can be fitted very well with two peaks that have approximately the height and width of the peak for the spin - ground state measured along the @xmath52-axis .
this provides further confirmation that the broadened peak is indeed a superposition of two separate peaks .
[ [ section-10 ] ] for the second measurement
@xmath49 always points in a direction perpendicular to the nanowire and is rotated from the @xmath52 axis at @xmath64 ( the same field direction as for @xmath55 in fig 3(a ) to the @xmath51 axis at @xmath65 . along the @xmath52 axis ,
the zeeman splitting is again @xmath66 @xmath57ev .
the zeeman splitting decreases until it reaches @xmath67 @xmath57ev along the @xmath51-axis , which corresponds to a @xmath0-factor of @xmath68 . in fig .
[ fig3](e ) two line cuts from fig .
[ fig3](b ) along the @xmath52-axis ( red curve ) and the @xmath51-axis ( blue curve ) are presented along with the fitted curves that were used to calculate @xmath28 .
[ [ section-11 ] ] the third rotation plane is the @xmath51-@xmath50 plane , with @xmath69 pointing along the @xmath51-axis , and @xmath70 along the @xmath50-axis [ see fig .
[ fig3](c ) ] . at @xmath69
we measure a spin splitting of @xmath71 @xmath57ev , corresponding to a @xmath0-factor of @xmath72 .
rotation of @xmath49 towards the @xmath50 axis again results in a zeeman splitting of @xmath73 @xmath57ev , corresponding to @xmath74 .
[ [ section-12 ] ] combining the three rotation experiments , the zeeman splitting along each of the @xmath51 , @xmath52 , and @xmath50 axes is measured twice with consistent values across experiments for the effective @xmath0-factor .
thus we can summarize our findings in fig .
[ fig3](f ) .
[ [ section-13 ] ] let us now compare our findings with experimental and theoretical results from the literature .
an anisotropy of the effective @xmath0-factor has been measured in other systems like si nanowire metal - oxide - semiconductor field - effect transistors ( mosfets ) ( @xmath75 ) @xcite , inas nanowires ( @xmath76 ) @xcite , and insb nanowires ( @xmath77 ) @xcite , all an order of magnitude smaller than our findings of @xmath78 for rotations with respect to the nanowire axis .
also self - assembled sige islands on si have been used for studies on the anisotropy of the effective @xmath0-factor ( @xmath79 ) @xcite , a system that is similar to ours , but lacking the one - dimensional confinement as well as the pronounced strain of our nanowires .
none of the reported anisotropies has been attributed to tunable electric fields , where we find @xmath80 for rotation with respect to the @xmath81-field axis . for ge / si core - shell nanowires , hu _ et al . _
@xcite reported an effective @xmath0-factor of @xmath82 measured with the @xmath27-field aligned along the nanowire axis @xmath83 .
@xcite have measured @xmath1 for different transitions ranging from 1.6 to 2.2 , @xmath49 was here aligned perpendicular to the nanowire .
both values are consistent with our measurements .
[ [ section-14 ] ] a @xmath0-factor anisotropy can in principle be related to the crystal direction @xcite . in our probably @xmath84110@xmath85-oriented device ,
we have observed a different anisotropy for holes states most probably originating from higher subbands , while we have observed qualitatively the same anisotropy in a second device tuned to the few - hole regime .
therefore we have strong evidence that the @xmath0-factor anisotropy observed here is rather related to an electric - field induced mixing between the lowest - lying subbands as discussed in the following paragraph .
[ [ section-15 ] ] maier _ et al . _
@xcite theoretically investigated the @xmath0-factor in ge - si core - shell nanowires .
they assumed elongated quantum dots , i.e. @xmath86 , which is very well fulfilled in our device , where @xmath87 nm and @xmath88 nm .
they predicted the @xmath0-factor to be highly anisotropic , with @xmath0 indeed being quenched along the nanowire axis , and a maximum @xmath0-factor perpendicular to the nanowire .
this is in excellent agreement with our measurements .
moreover , they predicted a lower @xmath0-factor at finite electric fields .
in particular , their calculations showed a more effective diminishment for @xmath89 than for @xmath90 .
also this agrees well with our findings .
show that this tunability of the @xmath0-factor with electric fields is caused by the effective coupling of different subbands through these electric fields and the mixed heavy - hole light - hole nature of the individual subbands .
the combination leads to a very pronounced spin - orbit interaction ( soi ) introduced as the ` direct rashba spin - orbit interaction ' @xcite because of its resemblance of the standard rashba soi and the fact that it is a leading - order process not suppressed by the band gap and thus expected to be 10 - 100 times stronger than the standard rashba soi for geometries similar to our device .
[ [ section-16 ] ] also quantitatively our measurements agree very well with the predictions regarding the @xmath0-factor quenching along the nanowire axis .
the @xmath0-factor suppression by the electric field is less pronounced than the calculations .
this can be explained by differences in the exact geometry of the wires , the quantum dot not being in the single - hole regime , and the fact that our device is operated at significantly higher electric fields than assumed by maier _
et al_. [ [ section-17 ] ] let us now briefly discuss the implications of these results for quantum computation applications .
a main obstacle for spin - based qubits is the fast coherent manipulation of the spin state . in principle
this can be done with a pulsed high - frequency ( hf ) magnetic field , but this is technologically very challenging .
our results imply that it is not only possible to use pulsed hf electric fields as also used in other systems with significant soi @xcite , but with continuous hf electric fields while tuning the qubit in and out of resonance by changing the static electric field , e.g. , through a combination of top and bottom gates .
in conclusion , we have demonstrated control over the hole occupancy in a ge - si core - shell nanowire quantum dot over 60 charge transitions .
the effective @xmath0-factor has been found to be highly anisotropic with respect not only to the nanowire axis but also the electric - field direction . in particular
we have found excellent qualitative agreement between our measurements and theoretical calculations @xcite .
this opens the way to controlled manipulation of the spin - state with a continuous high - frequency electric field , a major technological advancement .
we thank christoph kloeffel , daniel loss , and wilfred van der wiel for fruitful discussions and careful reading of the manuscript .
we acknowledge technical support by sergey amitonov , paul - christiaan spruijtenburg , and hans mertens .
f.a.z . acknowledges financial support through the ec seventh framework programme ( fp7-ict ) initiative under project siam no .
610637 , and from the foundation for fundamental research on matter ( fom ) , which is part of the netherlands organization for scientific research ( nwo ) .
e.p.a.m.b . acknowledges financial support through the ec seventh framework programme ( fp7-ict ) initiative under project sispin no .
323841 .
44ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _
_ ( , ) http://lib.semi.ac.cn:8080/tsh/dzzy/wsqk/science/vol270/270-255.pdf [ * * , ( ) ]
link:\doibase 10.1038/nature08812 [ * * , ( ) ] link:\doibase 10.1007/bf01886518 [ * * , ( ) ] link:\doibase 10.1038/414883a [ * * , ( ) ] link:\doibase 10.1103/physrevx.2.031007 [ * * , ( ) ] link:\doibase 10.1038/nature15263 [ * * , ( ) ] link:\doibase 10.1103/physreva.57.120 [ * * , ( ) ] link:\doibase 10.1038/30156 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.87.703 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.85.961 [ * * , ( ) ] link:\doibase 10.1021/cr400261y [ * * , ( ) ] link:\doibase 10.1143/jjap.42.6248 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/nnano.2014.211 [ * * , ( ) ] http://dx.doi.org/10.1038/nnano.2014.216 10.1038/nnano.2014.216 http://www.nature.com/nnano/journal/v9/n12/abs/nnano.2014.216.html{#}supplementary-information [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.073313 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.155323 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.195314 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.161305 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.241405 [ * * , ( ) ] link:\doibase 10.1038/nnano.2006.140 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.186802 [ * * , ( ) ] link:\doibase 10.1038/nnano.2011.234 [ * * , ( ) ] link:\doibase 10.1021/nl501242b [ * * , ( ) ] link:\doibase 10.1038/nnano.2007.302 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.216806 [ * * , ( ) ] @noop * * ( ) link:\doibase 10.1103/physrevlett.81.5225 [ * * , ( ) ] http://www.ncbi.nlm.nih.gov/pubmed/16178262 [ * * , ( ) ] in @noop _ _ ( , ) pp .
[ * * , ( ) ] link:\doibase 10.1088/0034 - 4885/64/6/201 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.79.1217 [ * * , ( ) ] link:\doibase 10.1088/0957 - 4484/21/27/274018 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.77.342 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.115438 [ * * , ( ) ] http://dx.doi.org/10.1021/acs.nanolett.5b02920 [ ( ) ] link:\doibase 10.1103/physrevlett.107.176811 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/nnano.2010.84 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.035315 [ * * , ( ) ] , link:\doibase 10.1126/science.1148092 [ * * , ( ) ] link:\doibase 10.1038/nature09682 [ * * , ( ) ] | we present angle - dependent measurements of the effective @xmath0-factor @xmath1 in a ge - si core - shell nanowire quantum dot .
@xmath1 is found to be maximum when the magnetic field is pointing perpendicular to both the nanowire and the electric field induced by local gates .
alignment of the magnetic field with the electric field reduces @xmath1 significantly .
@xmath1 is almost completely quenched when the magnetic field is aligned with the nanowire axis .
these findings confirm recent calculations , where the obtained anisotropy is attributed to a rashba - type spin - orbit interaction induced by heavy - hole light - hole mixing . in principle , this facilitates manipulation of spin - orbit qubits by means of a continuous high - frequency electric field . |
first - principles molecular dynamics ( fpmd ) based on density functional theory ( dft ) is a well - established and highly - successful tool for studying reactions or processes of materials at the atomic scale . due to the increase of the computer power ,
the variety and complexity of the materials and phenomena investigated by fpmd simulations have been growing . however , the size of the systems modelled with fpmd simulations have remained limited to systems of a few hundred atoms in most cases , because the computational cost of standard dft methods grow rapidly , proportional to the cube of the number of atoms @xmath0 in the system .
there are many demands to enlarge the system size in diverse fields , including computational physics , chemistry , materials science , biology and so on .
examples of these problems include chemical reactions at liquid / solid interfaces , various processes in complex biological systems , or the growth mechanism of the nano - structured materials at atomic scale ; in all these cases , we have to treat systems containing many thousands or tens of thousands of atoms . in this respect , recent advances in developing linear - scaling or o(@xmath0 ) methods are encouraging@xcite .
there have been several demonstrations that efficient and reliable linear - scaling dft calculations are now available to calculate the electronic structure , total energy and atomic forces of very large systems , including up to millions of atoms@xcite .
although it is now possible to calculate the total energy and atomic forces of very large systems using o(@xmath0 ) dft methods@xcite , this does not guarantee that stable , efficient and accurate fpmd simulations are possible in practice@xcite . with conventional dft methods , there are two widely - used methods to achieve efficient fpmd simulations .
the first method is known as car - parrinello md ( cpmd)@xcite , where the propagation of electronic structure and the atomic positions are treated simultaneously by introducing a fictitious mass for the electronic degree of freedom .
this method is efficient , but the accuracy of the cpmd simulations can depend on the choice of the fictitious mass .
by contrast , born - oppenheimer md ( bomd ) deals with the electronic and nuclear problems separately . in this method ,
the electronic structure in the ground state is calculated at each set of atomic positions , usually by the optimization of the kohn - sham orbitals using an iterative method@xcite .
the bomd method thus needs more cpu time than cpmd for each md step , but the method is robust and stable , and allows us to adopt a longer time step . as a result ,
the cpu time needed for a whole md simulation is tractable because the number of force calculations is smaller ( the cost of calculating atomic forces is usually much higher than the update of the kohn - sham orbitals ) .
the main problem facing this method , however , is that both the stability and the reliability of the md depend on the accuracy of the calculated forces . if the optimization of the kohn - sham orbitals or the convergence of scf is not sufficiently accurate , the method suffers from a systematic energy drift in micro - canonical md simulations , due to the violation of time - reversal symmetry
this can also result in poor reliability for canonical md simulations .
recently this problem has been solved using the extended lagrangian born - oppenheimer md ( xl - bomd ) proposed by niklasson et al.@xcite , whose lagrangian includes auxiliary electronic degrees of freedom as dynamical variables to recover time - reversal symmetry in bomd simulations . they have shown that bomd simulations with long - term conservation of total energy are possible .
it is reasonable to expect that this recent progress with the xl - bomd method could be used with o(@xmath0 ) dft techniques .
however , since linear - scaling methods rely on the locality of the electronic structure and need to introduce approximations to utilize this locality , the accuracy of o(@xmath0 ) method may change during the md simulations , and it is not clear how accurate or stable md can be with o(@xmath0 ) dft methods .
although there have been already a few reports showing the examples of o(@xmath0 ) fpmd@xcite or o(@xmath0 ) self - consistent tight - binding@xcite simulations , they use different linear scaling techniques and the information about the accuracy , stability or efficiency of the simulations is rather limited . in this paper , we report the stability and accuracy of the xl - bomd simulations with a density matrix minimization ( dmm ) technique@xcite , which is one of the most common o(@xmath0 ) dft methods .
the accuracy of the forces calculated by the dmm method is high because the method satisfies the variational principle@xcite . to our knowledge
, the combination of dmm and xl - bomd methods has not been investigated so far . since the original xl - bomd methods assume the orthogonality of the basis functions , the formulation of the method for non - orthogonal basis functions is also presented .
we show that the xl - bomd method can be introduced to our linear - scaling dft code conquest@xcite , and report that robust , accurate and efficient fpmd is possible with the dmm method .
the reliable calculation conditions for the accurate o(@xmath0 ) fpmd are also investigated . in the end , we demonstrate that it is now possible to do actual fpmd simulations of a 32,768-atom system . the rest of the paper is organized as follows . in the next section ,
we show the methods and algorithms which are used or introduced in this work . in sec . 3 , we report the results of the test calculations , investigation of the reliable calculation conditions , and some examples of the fpmd simulations , including that of 32,768-atom system . finally , concluding remarks are given in sec .
the implementation of molecular dynamics and all calculations were performed with , which is a linear scaling , pseudopotential dft code@xcite . in order to achieve linear - scaling both in computational time and memory consumption
, the code works with the density matrix @xmath1 rather than the wavefunctions , because of its spatial locality : @xmath2 assumes that the density matrix can be described in a separable form : @xmath3 where @xmath4 is a non - orthogonal localised orbital called a _ support function _ , which is only non - zero within a sphere centred at atom @xmath5 , and @xmath6 denotes the support functions of a given atom .
support functions are represented in terms of localised basis functions ; in we can use systematically improvable basis functions ( b - splines@xcite ) or pseudoatomic - orbitals ( paos)@xcite .
@xmath7 is the density matrix represented in terms of the support functions .
we find the electronic ground state by minimising the total energy with respect to the density matrix , while imposing the correct electron number and weak idempotency .
the complete minimisation combines two algorithms@xcite : mcweeny s iterative purification@xcite ; and the auxiliary density matrix ( adm ) method proposed by li , nunes and vanderbilt@xcite .
an initial density matrix is generated starting from the hamiltonian ( scaled to have the right trace and eigenvalue spectrum ) , using the extension of mcweeny s approach for the canonical case@xcite . in the adm method ,
weak idempotenty is imposed by expressing the density matrix @xmath8 in terms of an auxiliary density matrix @xmath9 , and the overlap @xmath10 between support functions , @xmath11 , as @xmath12 the spatial truncation on the density matrix is actually imposed on the matrix @xmath9 , such that @xmath13 once @xmath14 ; this is referred as _ l - range _ in the rest of this report .
the mcweeny s procedure is applied until the energy rises ( a sign of reaching truncation error@xcite ) .
the @xmath9 matrix is then passed into the adm , which is a variational approach , using a modified pulay / broyden algorithm .
this part of the minimisation is referred to as dmm ( density matrix minimisation ) .
+ in this paper , we represent each support function @xmath15 in terms of one pao ( the approach given here is easily extended to the case where we also minimise the energy with respect to the paos , which will be presented in a future publication ) . in running bomd simulations with the dmm method
, we have two ways to optimize the electronic structure .
one is to initialise the density matrix using mcweeny s procedure at every atom movement .
the initial charge density is made from the superposition of the atomic charge density .
since the hamiltonian and overlap matrices are recalculated after the movement of atoms , the initial matrix @xmath9 and subsequent optimised @xmath9 matrix will maintain time - reversal symmetry ; we hence expect that bomd simulations should conserve energy and be stable .
the other method is to utilize the auxiliary density matrix found from the previous md step to start the present dmm . in this case , the initial charge density is made from the initial density matrix @xmath9 prepared for the present dmm , and is updated with the @xmath9 matrix simultaneously during the dmm .
this method will be more efficient , but will break time reversal symmetry , potentially leading to unphysical energy drift . to maintain time - reversal symmetry while keeping the efficiency of density matrix re - use , we will use the extended lagrangian scheme proposed by niklasson et al.@xcite we introduce an auxiliary degree of freedom to the born - oppenheimer lagrangian @xmath16 , @xmath17 .
this is associated with @xmath18 , rather than @xmath9 to maintain the correct non - orthogonal metric@xcite .
@xmath19 - \frac{\mu}{2 } { \rm tr}\left [ \dot{x}^2 \right ] - % \frac{\mu \omega^2}{2 } \text{tr}\left [ \left ( ls - x \right)^2 \right ] \frac{\mu \omega^2}{2 } { \rm tr}\left [ \left ( ls - x \right)^2 \right ] \label{eq : xlbomd}\end{aligned}\ ] ] @xmath17 is a sparse matrix with the range of the matrix @xmath18 , @xmath20 the fictitious electronic mass , and @xmath21 is the curvature of the electronic harmonic potential . as in the original xl - bomd method , if we take the limit @xmath22 , @xmath16 becomes @xmath23 and we have equations of motion for nuclear positions and @xmath17 .
if we apply the time - reversible verlet scheme to calculate @xmath17 using the equation of motion , we have @xmath24 which shows that @xmath25 is time reversible and evolves in a harmonic potential centered around the ground - state @xmath26 .
this also implies that a good initial guess for the @xmath9-matrix , which will obey time reversal symmetry , can be calculated by multiplying @xmath17 and @xmath27 ( in , the sparse approximate inverse @xmath10 is computed using hotelling s method@xcite ) .
even though the time - reversibility is maintained during propagation , the @xmath17-matrix tends to move away from the harmonic centre over time . as a result ,
the number of iterations to reach the ground state at each md step gradually increases in the course of a simulation .
to remove accumulated numerical errors , we add a dissipative force to the propagation of @xmath17 following ref .
@xcite , @xmath28 here , the parameters @xmath29 , @xmath6 and @xmath30 are determined so that the dissipation term does not significantly break the time - reversal symmetry ; we use the values in ref .
the amount of dissipation is controlled with a single parameter @xmath31 , which determines the order of the polynomial , and we use @xmath32 in this paper . with the dissipation term , we can keep the kinetic energy of @xmath33 small and thus the @xmath25 is kept close to the present ground - state @xmath26 . as a result , we expect that the number of dmm steps would be reduced .
the effect of the dissipation term in the practical calculations is reported and discussed in sec .
[ sec : effect - schem - line ] .
we will present our tests of the implementation of o(@xmath0 ) dft in different stages : first , considering approaches that permit energy conservation and their efficiencies ; second , the effect of density matrix range and minimisation tolerance ; and finally presenting practical calculations .
first , we explore methods to achieve energy conservation with o(@xmath0 ) dft and bomd .
we monitor the born - oppenheimer total energy @xmath34 , which is defined as a sum of the ionic kinetic energy @xmath35 and the dft total energy @xmath36 . in the micro - canonical ensemble ,
@xmath34 should be constant and so is a good indicator to judge whether a simulation is accurately carried out . in what follows ,
@xmath37 is set to 16 bohr ( the effect of changing this is explored below in sec .
[ sec : calc - cond - bmmathc ] ) .
a single-@xmath38 ( sz ) pao is used to represent support functions@xcite , and the local density approximation ( lda ) is adopted to the exchange - correlation functional with the parametrization by perdew and zunger@xcite .
md simulations are conducted via the velocity - verlet integrator@xcite with a time step being 0.5 femtoseconds ( fs ) in a micro - canonical ensemble ; the initial velocities are given randomly so that the maxwell - boltzmann distribution holds with 300 k. the same conditions are set in other simulations in this report unless otherwise specified .
for a first run , we perform bomd on a 64-atom crystalline silicon with density - matrix initialisation by mcweeny step , so that the density matrix is initialised from the hamiltonian at each md step .
the density matrix is updated until the residual in the dmm step , @xmath39 with @xmath40 , becomes smaller than a given tolerance @xmath41 .
here , @xmath42 is defined as @xmath43 , the number of electrons is @xmath44 and the lagrange multiplier @xmath20 is used to keep the electron number fixed .
the profile of the born - oppenheimer energy @xmath34 is shown by a solid line in fig .
[ fig : eprofiles1 ] , and it is clear that the total energy is conserved even for a low dmm tolerance , @xmath45 .
this simulation is , however , time - consuming since the electronic structure is calculated from scratch at every atom movement .
we now consider the re - use of the optimised @xmath9-matrix from the previous md step . when the optimised @xmath9-matrix from the previous step is used to initialise the density matrix , the cpu time is significantly reduced .
however , as well as this improved efficiency , we observe a significant and undesired energy drift , as shown in fig .
[ fig : eprofiles1 ] with symbols . here
, we consider various values for the tolerance applied to the density matrix minimisation , @xmath41 .
the energy drift can be reduced by using a tighter @xmath41 , but tighter tolerance rapidly increases the number of dmm iterations and the computational cost . in order to solve the problem of both accuracy and efficiency , we carry out xl - bomd with a dmm tolerance , @xmath45 .
its @xmath34 profile ( crosses in fig . [
fig : eprofiles2 ] ) agrees extremely well with that of bomd with density - matrix initialisation by means of mcweeny step ( solid line in fig .
[ fig : eprofiles2 ] ) , implying that xl - bomd yields a well conserved energy with lower computational requirements . however , there is a problem in the efficiency with the method , as explained in sec .
2 . after a long simulation time , the number of dmm iterations required to find the ground state density matrix grows .
this growing cost can be suppressed by applying a dissipative force , and the resulting @xmath34 profile , circles in fig .
[ fig : eprofiles2 ] , is still very close to the profile without dissipation .
a comparison of the required number of dmm iterations at each step for these different approaches is shown in table .
[ table : dmmcost ] .
the reduction of computational cost by using xl - bomd with dissipation is significant . besides the smaller number of dmm iterations ,
the method does not require mcweeny steps which typically include a few tens of matrix multiplications for purification ( as in eq .
[ eq : purification ] ) . taking values 1.6@xmath46 ( circle ) , 1.6@xmath47 ( square ) and 1.6@xmath48 ( triangle).,width=340,height=283 ] 1.6@xmath46 , and different initialisations : mcweeny initialisation at each md step ( solid line ) ; xl - bomd with dissipative force ( circles ) ; and xl - bomd without dissipative force ( crosses).,width=340,height=283 ] .dmm iterations required for each md step ; bomd with density - matrix initialisation by mcweeny step , and xl - bomd in the presence ( y ) and absence ( n ) of a dissipative force .
averages taken over 1,000 steps . [ cols="<,>,>,>,>",options="header " , ] in the last section , we showed that xl - bomd simulation with dmm method was robust and efficient using one typical set of calculation conditions .
we now investigate how md results are affected by the parameters which control the accuracy and efficiency of linear - scaling dft calculations with dmm .
there are two key parameters : the range applied to the auxiliary density matrix ( @xmath9-range or @xmath49 ) ; and the tolerance applied to the minimisation , @xmath41 . in both tests ,
xl - bomd is performed on a 64-atom crystalline silicon with a dissipative force .
the key approximation made in linear scaling is to localize the density matrix , and it is vital to examine how the conservation of energy , @xmath50 , is affected by the range , the parameter @xmath49 . we have tested three values of @xmath49 : 13 , 16 and 20 bohr , each with a strict dmm tolerance , @xmath51 .
simulations are performed for 2ps ( or 4,000 steps ) .
the time - averaged bo total energies are very different for the three ranges ( -106.830 ev / atom , -106.930 ev / atom and -106.987 ev / atom ) simply because the potential energy is lower for longer ranges .
we plot the instantaneous energy fluctuations for the three ranges in fig .
[ fig : rldep ] ; the time - averaged absolute fluctuations , @xmath52 , are 0.24 mev / atom , 0.15 mev / atom and 0.075 mev / atom .
these fluctuations become smaller as the range increases , and no energy drift is seen for any of the ranges . during md
, atoms will move in and out of the density matrix range , and we might expect to see energy drift and deviations at small ranges , compared to long ranges . the lack of drift is extremely encouraging , showing that the xl - bomd approach is well - suited both to orthogonal and non - orthogonal basis sets , and linear scaling approaches .
it is also encouraging that , for relatively modest ranges , good accuracy is seen .
we plot the kinetic energy of the ions in fig .
[ fig : trldep ] , which shows that , while there is a deviation for 13 bohr range after 2 - 300 femtoseconds , the 16 bohr range profile is extremely close to the 20 bohr range profile , and therefore this range can be used with confidence over at least 1ps@xcite . values ; 13 bohr ( top ) , 16 bohr ( middle ) and 20 bohr ( bottom).,width=264,height=377 ] values ; 13 bohr , 16 bohr and 20 bohr.,width=377,height=340 ] the other main parameter which must be tested for its effect on the accuracy and stability of linear - scaling md is the dmm tolerance @xmath41 . from the results of the last section , we fix the @xmath9-range to 16 bohr and apply a wide range of @xmath41 from @xmath53 to @xmath54 ; we assume that the results with @xmath55 can be regarded as reference values .
we plot the bo total energy over a set of 1ps simulations in fig .
[ fig : e ] .
first we note that @xmath50 with @xmath56 ( circles ) is not fully converged , and is higher in energy by about 2 mev / atom than the converged one . however , as we see in fig .
[ fig : t ] , the kinetic energy profile shows almost perfect agreement with the reference profile of @xmath57 , indicating that forces and hence the ionic kinetic energy converge faster than the total energy . note that we can achieve good convergence for both @xmath50 and @xmath35 with @xmath58 . in addition , the @xmath50 profiles are close to each other for all tolerances except for a constant shift of the energy .
given the fast convergence in ionic kinetic energy , and the similarities in @xmath50 profiles , the calculated trajectories will be very close .
hence , the dmm tolerance of @xmath53 will give an accurate simulation .
values : circles , squares and triangles denote the profiles given by @xmath59,@xmath60 and @xmath61 , respectively .
the solid line shows the profile calculated with @xmath55.,width=377,height=340 ] values : circles , crosses and triangles denote the profiles given by @xmath59,@xmath60 and @xmath61 , respectively .
a solid line shows the profile calculated with @xmath55.,width=377,height=340 ] having investigated the effects of the parameters on simulations , we now present applications of the code to practical md simulations using a double-@xmath38 + polarisation ( dzp ) pao basis , which is commonly used as a converged basis set .
three different systems are presented : 8-atom crystalline silicon ; 32 water molecules ; and a 32,768-atom crystalline silicon cell to demonstrate application to a large system .
for all systems , @xmath9 range is set to 16 bohr .
the initial temperature is 300k and the time step of the simulation is 0.5 femtosecond , unless stated otherwise .
our first example is a small sample of crystalline silicon .
we show the bo energy , @xmath34 , calculated with @xmath41 set to @xmath63 in figure [ fig : boesi8dzp ] .
this plot demonstrates excellent energy conservation , though there are large fluctuations for the first few femtoseconds .
this initial large amplitude is probably caused by the system not being in an equilibrium state at the start of the simulation , and is not a significant problem . in the inset of fig .
[ fig : boesi8dzp ] , we also show the result of a simulation where we set @xmath64 in eq .
[ eq : xlbomd ] or eq .
[ eq : trbomd ] ( by supposing that @xmath9 is expressed using orthogonal basis sets ) . by comparing the result ( dotted line ) with the correct one ( solid line )
, we see that the former energy exhibits significant drift , while the latter one is well behaved .
this indicates the importance of the metric used for non - orthogonal basis functions .
we performed another md simulation on the same system using the same calculation conditions except that the time step is increased to 2.0 femtoseconds ( this result is shown in supporting information ) .
we see that the large oscillations observed in the early stage are even larger and the time to suppress them becomes longer .
however , it is found that the time evolution of the kinetic energy of si atoms is almost the same between these two md simulations .
this result means that time step of 2.0 femtoseconds can be used to do reliable md simulations on this silicon system .
it is encouraging that we can use such a large time step for o(@xmath0 ) fpmd simulations . here , we should note that a new efficient method which does not require self - consistent field optimization has been recently proposed@xcite within the xl - bomd framework . introducing such new techniques to the conquest code may even further accelerate the speed of o(@xmath0 ) fpmd simulations .
we take a box of water made up of 32 molecules as our next target .
water is important as a system in its own right , as well as forming the environment for most biological simulations .
we performed linear scaling xl - bomd with @xmath41 set to @xmath53 , and using the gga - pbe functional@xcite for exchange - correlation , as it describes water better than lda ( we note that the convergence with respect to the grid may be slower for gga ) . in fig .
[ fig : boewater32 ] we show the total energy and the ionic and potential energies , @xmath35 and @xmath36 .
we see that there are no initial large oscillations unlike the silicon case , and that @xmath35 appears to keep increasing during the run .
this behavior is seen because we used the micro - canonical ensemble and the simulation time is too short to reach a steady state . even in such a situation showing fast changes , the averaged fluctuations are @xmath65 mev atom@xmath66 and good energy conservation is found without any energy drift .
these results clearly demonstrate that there are no effects from the grid on the stability and accuracy of md . , and red dashed and blue dotted lines characterise its components , @xmath67 and @xmath35 , respectively ( color online).,width=377,height=340 ] finally ,
as a demonstration of the scalability of our approach , we apply our algorithm to a large system : a 32,768 atom crystalline silicon sample .
the time evolution of the bo and potential energy are shown in fig .
[ fig : si32768 ] .
this is our first substantial attempt to examine the stability of an md simulation on massive systems .
this simulation is conducted under the same conditions as those in the 8-atom system in fig .
[ fig : boesi8dzp ] , except that the time step is set to 2.0 femtoseconds .
we observe some similarities to the previous results . in the early stage of a simulation ,
large oscillations are seen in the energy , but are gradually suppressed during the md run .
the time to suppress the large oscillations is longer than the one in fig .
[ fig : boesi8dzp ] , but similar to the case using 2.0 femtoseconds for 8-atom crystalline silicon .
( see fig .
s1 in the supporting information . )
thus , this comes from the difference of the time step , not from the system size .
the energy fluctuations after 100 femtoseconds is smaller than 1 mev atom@xmath66 and there is no energy drift , as shown in fig .
[ fig : si32768 ] .
although we only carried out a short md run , the simulation is expected to remain stable in longer cases without any drift , based on the preceding example of an 8-atom system . moreover
, our approach is expected to ensure great stability on even larger and more complex systems by virtue of high computational efficiency realised in the code@xcite .
the average time required for 1 md step was 1085 seconds using 1024 cpus ( with 16 cores / cpu ) of fujitsu fx10 .
we are confident that this wall - clock time will be greatly reduced in the near future , because the code is not optimized at present , especially for the part updating @xmath17 matrices ; for simplicity , we now use a simple disk - based i / o procedure to read and write matrix elements of @xmath17s in the previous steps and this part will be improved in the near future .
in this paper , we have investigated the efficiency , stability and accuracy of md simulations with the code using the dmm linear scaling approach , to demonstrate large - scale o(@xmath0 ) fpmd simulations in practice .
we first found that accurate md simulations can be performed by calculating the electronic structure from scratch using the mcweeny procedure at every atomic configuration .
this gives excellent energy conservation even with a rather rough tolerance on the convergence of the density matrix minimisation .
however , this method requires considerable computational time because it initialises the density matrix at each step , rather than re - utilising it .
direct re - use of the optimised density matrix reduces computational costs considerably , but causes an unphysical energy drift . in order to achieve both accuracy and efficiency ,
we have introduced the xl - bomd formalism , proposed by niklasson et al .
, to the code , formulated for non - orthogonal basis functions .
the application of xl - bomd turned out to be much more efficient than the mcweeny approach , especially with the aid of a dissipative force , and good energy conservation was observed .
we studied how calculated results depend on two parameters specific to the dmm method : the cutoff range of the @xmath9 matrix and the tolerance applied to the density matrix minimisation .
we found that larger @xmath9-ranges resulted in smaller energy fluctuations .
moreover , we found that the md trajectories from runs with different ranges are almost identical , even when the born - oppenhimer total energy was not fully converged .
this indicates that , as is well known , the atomic forces converge much faster than the total energy with respect to these two parameters . as practical applications
, we treated crystalline silicon with lda and liquid water with gga . for the crystalline silicon , we demonstrated that accurate md runs can be performed on very large systems , in this case containing 32,768 atoms , but scalable to significantly larger systems .
the algorithms presented in this paper are applicable to a canonical ensemble and thus will open doors to more practical calculations on very large and complex systems .
the authors thank lianheng tong and takao otsuka for fruitful discussions .
this work is partly supported by kakenhi projects by mext ( no . 22104005 ) and jsps ( no . 26610120 and 26246021 ) , japan .
the support from the strategic programs for innovative research ( spire ) and the computational materials science initiative ( cmsi ) is also acknowledged .
calculations were performed using the numerical materials simulator at nims , tsukuba , japan , k computer at riken advanced institute for computational science(aics ) , kobe , japan , and the fujitsu fx10 system at the information technology center , university of tokyo , tokyo , japan .
we present the comparison of the profiles of born - oppenheimer and kinetic energy using different timesteps , 0.5 fs and 2.0 fs and show that the trajectory of the ions are almost the same .
the result of crystalline silicon made of 4096 atoms using the initial temperature of 600k is also provided and we see that the energy convergence is excellent .
bowler , d. r. ; miyazaki , t. _ rep .
phys . _ * 2012 * , _ 75 _ , 036503 bowler , d. r. ; miyazaki , t. _ j. phys . : condens
. matter _ * 2010 * , _ 22 _ , 074207 arita , m. ; arapan , s. ; bowler , d. r. ; miyazaki , t. _ j. adv . simulat .
eng . _ * 2014 * , in press soler , j. m. ; artacho , e. ; gale , j. d. ; garca , a. ; junquera , j. ; ordejn , p. ; snchez - portal , d. _ j. phys.:condens . matter _ * 2002 * , _ 14 _ , 24752779 miyazaki , t. ; bowler , d. r. ; choudhury , r. ; gillan , m. j. _ j. chem .
phys . _ * 2004 * , _ 121 _ , 61866194 hine , n. d. m. ; robinson , m. ; haynes , p. d. ; skylaris , c .- k . ; payne , m. c. ; mostofi , a. a. _ phys . rev .
b _ * 2011 * , _ 83 _ , 195102 osei - kuffuor , d. ; fattebert , j .- l . _ phys
* 2014 * , _ 112 _ , 046401 bowler , d. ; miyazaki , t. ; truflandier , l. a. ; gillan , m. j. comment on `` accurate and scalable o(n ) algorithm for first - principles molecular - dynamics computations on large parallel computers '' .
2014 , eprint arxiv:1402.6828 , http://arxiv.org/abs/1402.6828 ( accessed sep 22 , 2014 ) car , r. ; parrinello , m. _ phys .
lett . _ * 1985 * , _ 55 _ , 24712474 payne , m. c. ; teter , m. p. ; allan , d. c. ; arias , t. a. ; joannopoulos , j. d. _ rev .
phys . _ * 1992 * , _ 64 _ , 10451097 niklasson , a. m. n. ; tymczak , c. j. ; challacombe , m. _ phys .
_ * 2006 * , _ 97 _ , 123001 niklasson , a. m. n. _ phys . rev .
lett . _ * 2008 * , _ 100 _ , 123004 niklasson , a. m. n. ; steneteg , p. ; bock , n. _ j. chem .
phys . _ * 2011 * , _ 135 _ , 164111 ohwaki , t. ; otani , m. ; ikeshoji , t. ; ozaki , t. _ j. chem .
phys . _ * 2012 * , _ 136 _ , 134101 ohwaki , t. ; otani , m. ; ozaki , t. _ j. chem . phys . _ * 2014 * , _ 140 _ , 244105 shimojo , f. ; hattori , s. ; kalia , r. k. ; kunaseth , m. ; mou , w. ; nakano , a. ; nomura , k .- i . ; ohmura , s. ; rajak , p. ; shimamura , k. ; vashishta , p. _ j. chem
. phys . _ * 2014 * , _ 140 _ , 18a529 tsuchida , e. _ j. phys . : condens
. matter _ * 2008 * , _ 20 _ , 294212 cawkwell , m. j. ; niklasson , a. m. n. _ j. chem .
phys . _ * 2012 * , _ 137 _ , 134105 li , x .- p . ; nunes , r. w. ; vanderbilt , d. _ phys .
b _ * 1993 * , _ 47 _ , 1089110894 hernndez , e. ; gillan , m. j. _ phys . rev . b _ * 1995 * , _ 51 _ , 1015710160 hernndez , e. ; gillan , m. j. ; goringe , c. m. _ phys . rev .
b _ * 1996 * , _ 53 _ , 71477157 bowler , d. r. ; miyazaki , t. ; gillan , m. j. _ j. phys.:condens . matter _ * 2002 * , _ 14 _ , 27812798 bowler , d. r. ; choudhury , r. ; gillan , m. j. ; miyazaki , t. _ phys .
( b ) _ * 2006 * , _ 243 _ , 9891000 hernndez , e. ; gillan , m. j. ; goringe , c. m. _ phys .
b _ * 1997 * , _ 55 _ , 1348513493 torralba , a. s. ; todorovi , ; m. , v. , brzdov ; choudhury , r. ; miyazaki , t. ; gillan , m. j. ; bowler , d. r. _ j. phys.:condens . matter _ * 2008 * , _ 20 _ , 294206 bowler , d. r. ; gillan , m. j. _ comp . phys .
commun . _ * 1999 * , _ 120 _ , 95108 mcweeny , r. _ rev .
phys . _ * 1960 * , _ 32 _ , 335369 palser , a. h. r. ; manolopoulos , d. e. _ phys .
b _ * 1998 * , _ 58 _ , 1270412711 as each support function is represented by only one pao , we need not consider the time evolution of @xmath10 . press , w. h. ; flannery , b. p. ; teukolsky , s. a. ; vetterling , w. t. _ numerical recipes in fortran : the art of scientific computing _ , 2nd ed .
; cambridge university press : cambridge , uk , 1992 niklasson , a. m. n. ; steneteg , p. ; odell , a. ; bock , n. ; challacombe , m. ; tymczak , c. j. ; holmstrm , e. ; zheng , g. ; weber , v. _ j. chem .
_ * 2009 * , _ 130 _ , 214109 this does not affect the results and larger basis sets are used below perdew , j. p. ; zunger , a. _ phys .
b _ * 1981 * , _ 23 _ , 50485079 verlet , l. _ phys . rev . _ * 1967 * , _ 159 _ , 98103 this does not mean the result of md by using @xmath37=13bohr is incorrect . the different trajectory may simply come from infinitesimal differences in forces of a chaotic system .
souvatzis , p. ; niklasson , a. m. n. _ the journal of chemical physics _
_ 140 _ , 044117 perdew , j. p. ; burke , k. ; ernzerhof , m. _ phys . rev . lett . _ * 1996 * , _ 77 _ , 38653868 torralba , a. s. ; bowler , d. r. ; miyazaki , t. ; gillan , m. j. _ j. chem .
theory comput . _ * 2009 * , _ 5 _ , 14991505 miyazaki , t. ultra - large - scale first - principles calculations by the k computer .
nims now vol .
11(9 ) , 6 , 2013 ; http://www.nims.go.jp/eng/publicity/nimsnow/2013/vol11_09.html ( accessed sep 22 , 2014 ) | the recent progress of linear - scaling or o(@xmath0 ) methods in the density functional theory ( dft ) is remarkable . given this , we might expect that first - principles molecular dynamics ( fpmd ) simulations based on dft could treat more realistic and complex systems using the o(@xmath0 ) technique .
however , very few examples of o(@xmath0 ) fpmd simulations exist so far and the information for the accuracy or reliability of the simulations is very limited . in this paper , we show that efficient and robust o(@xmath0 ) fpmd simulations are now possible by the combination of the extended lagrangian born - oppenheimer molecular dynamics method , which was recently proposed by niklasson _ et al _ ( _ phys
. rev .
lett .
_ 100 , 123004 ( 2008 ) ) , and the density matrix method as an o(@xmath0 ) technique . using our linear - scaling dft code conquest , we investigate the reliable calculation conditions for the accurate o(@xmath0 ) fpmd and
demonstrate that we are now able to do practical , reliable self - consistent fpmd simulation of a very large system containing 32,768 atoms .
some journals require a graphical entry for the table of contents .
this should be laid out `` print ready '' so that the sizing of the text is correct .
inside the ` tocentry ` environment , the font used is helvetica 8pt , as required by _ journal of the american chemical society_. the surrounding frame is 9 cm by 3.5 cm , which is the maximum permitted for _ journal of the american chemical society _ graphical table of content entries .
the box will not resize if the content is too big : instead it will overflow the edge of the box . this box and the associated title will always be printed on a separate page at the end of the document . |
in order to describe processes in nuclei involving large excursions from equilibrium , such as shape - coexistence or fission , we need to go beyond the simple harmonic oscillator picture underlying the random - phase approximation ( rpa ) .
one method is to use a method that self - consistently selects collective coordinates for large amplitude collective motion ( lacm ) .
recently we have analysed properties of collective motion in a pairing model hamiltonian with level crossings @xcite , where we used a version of a theory proposed by klein et al @xcite . to determine the collective coordinates .
we have shown that the system automatically selects either diabatic or adiabatic collective surfaces according to the strength of the pairing interaction .
the model used in that paper was rather artificial , since it involved a macroscopic system coupling to a few fermionic degrees of freedom .
we felt it desirable to test the theory for a fully microscopic hamiltonian which is able to describe systems from vibrational nuclei to well - deformed nuclei . in this paper
, we investigate the collective motion in a microscopic model which describes a system of nucleons interacting through a simplified version of the pairing plus quadruple force @xcite . although the hamiltonian has a very simple form , we shall see that the model can represent several of the interesting phenomena observed in real nuclei .
the theory of adiabatic large amplitude collective motion ( alacm , see reference @xcite for a complete description ) is applicable to a classical hamiltonian system which has kinetic terms only quadratic in momentum .
we thus have to start with a truncated hamiltonian @xmath0 where the mass tensor @xmath1 , in general , depends on the coordinates @xmath2 and is defined by truncation of the hamiltonian to second order .
collective coordinates @xmath3 and intrinsic ( non - collective ) coordinates @xmath4 which are approximately decoupled from each other , are assumed to be obtainable by making a point transformation ,
@xmath5 requiring approximate decoupling between the collective and non - collective coordinates , we find a set of coupled equations . in the case of a single collective coordinate ( @xmath6 )
, the equations can be written as @xcite @xmath7 the equations ( [ lhe_1 ] ) can be solved iteratively , starting from a stationary point . in principle , the procedure to find a collective path is to find successive points at which an eigenvector @xmath8 of the covariant rpa equation ( second equation ) satisfies the force condition ( first equation ) at the same time .
the model hamiltonian we discuss in this paper has @xmath9 symmetry.it has been originally developed to describe @xmath10 excitations in deformed nuclei @xcite .
it has also been used to test various methods for calculation of collective excitations @xcite .
the generalisation to multiple shells has also been formulated , in order to investigate shape - coexistence phenomena @xcite .
consider a single shell with angular momentum @xmath11 , and define the four basic operators , @xmath12 here @xmath13 if @xmath14 , and @xmath15 for the other values of @xmath16 , which mimics properties of the matrix elements of the quadrupole operator @xmath17 .
@xmath18 is assumed to be an even integer .
hereafter we call @xmath19 the quadrupole operator .
the four operators , @xmath20 , @xmath21 , @xmath22 and @xmath19 , close under commutation , and generate the lie algebra @xmath23 .
we now build a model hamiltonian with @xmath9 dynamical symmetry @xmath24 which contains both a monopole pairing and a quadrupole - quadrupole ( qq ) interaction .
the exact solution can be obtained by diagonalisation in an @xmath9 basis .
the behaviour of the model is based in the competition between the pairing force which tends to align quasi - spins and the qq force which tends to de - align them .
this can be made extremely clear in the limit of no pairing or no qq force , where the model can be solved exactly .
we can easily generalise the previous model into a multi-@xmath11-shell case , @xmath25 . for each @xmath11-shell , we take @xmath26 to be even , and we weigh the quadrupole operators by @xmath27 , the magnitude of quadrupole moment carried by the single - particle state @xmath11 .
we again use the pairing plus quadrupole hamiltonian , but we now add spherical single - particle energies , @xmath28
( numbers next to arrows ) in the two - shell @xmath9 model discussed in the text , as functions of quadrupole force strength @xmath29 .
the case ( a ) represents the almost harmonic motion for @xmath30 , ( b ) the flat potential for @xmath31 and ( c ) shape coexistence for @xmath32 . in each case
we compare an exact diagonalisation with the result of requantisation in a single collective variable . ]
one of the representative calculations we performed is for a two - shell problem .
both shells have @xmath33 , but we take @xmath34 and @xmath35 . for strong `` deformation ''
we expect that the shells mix strongly , and for weak qq force we expect a simple spectrum . in this case , as can be seen in figure 1 , we find that as we move from a harmonic spectrum to a more complex mixed situation , that we still keep reasonable agreement for both spectrum and transition strengths .
the values of the qq strength used were such that the middle case exactly captures the turn - over from single to double well .
in all these cases one collective coordinate appears to be enough .
there are some states in the double - well structure that are not well described by the current method , which is currently under investigation .
in conclusion we have shown that our method can deal with both anharmonic vibrations and shape coexistence . in future
we shall extend this research to the proton - neutron o(4 ) model , and to the real pairing - plus - quadrupole model .
9 nakatsukasa t and walet n r 1998 _ phys .
_ c * 57 * 1192 klein a , walet n r and do dang g 1991 _ ann .
_ ny*208 * 90 piepenbring a , silvestre - brac r and szymanski z 1980 _ nucl .
phys . _ * a348 * 77 matsuyanagi k 1982 _ prog .
phys . _ * 67
* 1441 matsuo m 1986 _ prog . theor . phys . _ * 76 * 372 suzuki t and mizobuchi y 1988 _ prog .
_ * 79 * 480 fukui t , matsuo m and matsuyanagi k 1991 _ prog .
phys . _ * 85 * 281 | we discuss the application of a theory of large - amplitude collective motion to a simple model mimicking the pairing - plus - quadrupole model of nuclear physics . |
it has long been recognized that the effective cosmological constant @xmath0 may have contributions from a four - form field @xmath1 , and that in such case @xmath2 may vary in space and time due to brane nucleation events .
this has led to various proposals for solving the cosmological constant problem , starting with the pioneering work of brown and teitelboim @xcite .
these authors considered a cosmological scenario where @xmath0 is initially very large and positive , due to a large @xmath3 term .
the additive constant @xmath4 in ( [ 1 ] ) is assumed to be negative , but not fine - tuned in any way , so its absolute value is expected to be of the order of some cut - off scale to the fourth power . during the cosmological evolution , @xmath0 is neutralized " through successive nucleation of closed 2-branes ( charged with respect to the form field ) , which decrease the value of @xmath1 , until eventually @xmath0 is relaxed down to the small observed value , @xmath5 .
one problem with the original scenario is that neutralization must proceed in very small steps , so that any initially large @xmath0 can be brought to @xmath5 without overshooting into negative values .
for that , the charge of the branes should be tiny , ensuring that @xmath6 at each step .
also , the nucleation rate must be very small , or else the present vacuum would quickly decay .
these two constraints make the relaxation process extremely slow on a cosmological time - scale .
meanwhile , ordinary matter in the universe is exponentially diluted by the quasi - de sitter expansion , resulting in a disappointing empty universe . recently , feng et al .
( fmsw ) @xcite have suggested that nucleation of coincident branes may offer a solution to the empty universe " problem .
their proposal can be summarized as follows . in the context of m - theory ,
a stack of @xmath7 coincident d - branes supports a number of low energy degrees of freedom , corresponding to a @xmath8 super yang - mills ( sym ) theory living on the world - sheet .
consequently , the nucleation rate of coincident branes should be accompanied by large degeneracy factors , and could in principle be enhanced with respect to the nucleation of single branes .
the charge of a stack of branes can be very large even if the individual charges are small , facilitating quick jumps from @xmath0 to @xmath5 . in this way
, neutralization might proceed very rapidly , perhaps in just a few multiple " steps of the right size .
finally , the stability of the present vacuum could be due to gravitational suppression of the nucleation rate @xcite .
fmsw argued , rather heuristically , that the nucleation rate of coincident branes should be enhanced by a factor of the form @xmath9 where @xmath10 is the entropy of the worldsheet sym fields .
this entropy was estimated through simple thermodynamic arguments , as @xmath11 where @xmath12 is the effective number of worldsheet field degrees of freedom , and @xmath13 is the size of the brane at the time of nucleation .
however it remained unclear in @xcite which temperature @xmath14 should be used for the worldsheet degrees of freedom .
brane nucleation takes place in an ambient de sitter ( ds ) space characterized by a gibbons - hawking temperature @xmath15 .
the region inside the closed brane has a smaller value of the effective cosmological constant , and is therefore characterized by a smaller temperature @xmath16 .
feng et al .
considered two alternative possibilities for the temperature of the worldsheet degrees of freedom : @xmath17 and @xmath18 .
the proposed enhancement of the nucleation rate and the resulting cosmological scenarios are quite different in both cases , and therefore it seems important to try and clarify the issue of which temperature is the relevant one . the purpose of this paper is to present a more formal derivation of the nucleation rate corresponding to multiple brane nucleation .
as we shall see , the temperature relevant for the worldsheet degrees of freedom is in fact determined by the internal geometry of the worldsheet @xcite . for the coleman - de luccia ( cdl )
instanton , this worldsheet is a 2 + 1 dimensional de sitter space of radius @xmath13 , and the corresponding temperature is @xmath19 . when substituted into the naive expression ( [ entro ] ) , this leads to @xmath20 , independent of @xmath13 ( and hence on the ambient ds temperatures ) . as we shall see , the actual result has a certain dependence on @xmath13 due to the anomalous infrared behaviour of light fields in the lower dimensional de sitter space , but this results only in a rather mild dependence on the ambient ds temperatures .
we shall also see that de sitter space allows for a static " instanton which may be quite relevant to the nucleation of coincident branes .
this is analogous to the instanton for thermal activation in flat space .
it has a higher euclidean action than the cdl solution , and hence ( ignoring the degeneracy factor ) it seems to represent a subdominant channel of decay .
however , we shall argue that , depending on the short distance behaviour of the interactions amongst the branes , the cdl instanton for coincident brane nucleation may simply not exist , and in this situation the static instanton may be the relevant one . in several respects , the static instanton appears to be better suited to the neutralization scenario proposed by feng et al . than the cdl one .
the paper is organized as follows . in section
ii we review different proposals for neutralization of @xmath0 via brane nucleation .
section iii contains a discussion of coincident branes in 4 spacetime dimensions .
these are obtained from dimensional reduction of type iia supergravity in ten dimensions . in the 4 dimensional picture ,
the gravitational and four - form forces are both repulsive . however , the two are exactly balanced by the attractive force mediated by the scalar dilaton . in section iv and
v we discuss the stabilization of the dilaton , which is required in a more realistic scenario .
after the dilaton aquires a mass , the remaining long range forces are repulsive , rendering the stack of coincident branes unstable , or metastable at best .
this has important implications , since the instanton for nucleation of coincident branes will only exist provided that some mechanism causes an attractive interbrane force at short distances .
section vi contains a description of the cdl instanton for nucleation of coincident branes , highlighting a few limiting cases of interest . in section vii
we discuss the corresponding degeneracy factor in the nucleation rate , and we show that its dependence on the ambient de sitter temperatures is rather mild .
we also include a heuristic interpretation of this results based on the observation that the relevant temperature for the worldsheet degrees of freedom is determined by the inverse of the radius of the instanton .
implications for the scenario of @xcite are briefly discussed .
section viii , is devoted to a study of the
static " instanton , where the worldsheet has the topology @xmath21 , and where the intrinsic temperature is comparable to @xmath22 . in this case
, the dependence of the nucleation rate on the ambient ds temperature is exponential .
coincident brane nucleation can be unsuppressed at large @xmath22 but strongly suppressed at present .
our conclusions are summarized in section ix .
some technical discussions are left to the appendices .
to conclude this introduction , a disclaimer may be useful . for most of the paper
, we shall work directly in four dimensions , and our discussion will be certainly less than rigorous from the string theory point of view . in particular , we shall model the degrees of freedom of a stack of @xmath7 coincident 2-branes by a weakly coupled @xmath8 gauge theory on the world - sheet .
this may or may not correspond to a true dimensional reduction from m - theory , but it should at least represent some of the broad features of the degeneracy factors .
in four dimensions , a four form can be written as @xmath24 where @xmath25 is the determinant of the metric and @xmath26 is the levi - civita symbol .
this field has no propagating degrees of freedom , since in the absence of sources the equation of motion @xmath27 implies that @xmath1 is a constant .
this simply gives a contribution to the effective cosmological constant which gets added to the true cosmological constant , or vacuum energy density @xmath4 , @xmath28 four - forms may couple to charged " 2-dimensional extended sources , or 2-branes , through a term of the form @xmath29 where the integral is over the worldsheet of the extended object and @xmath30 is the 3-form potential ( @xmath31 ) . in this case , @xmath1 changes by @xmath32 accross a brane of charge @xmath33 .
consequently , @xmath1 can decay through nucleation of closed spherical branes .
the process is analogous to pair creation in the presence of an electric field , and very similar to false vacuum decay in field theory @xcite .
the closed brane is the boundary of a newly formed
true " vacuum bubble , where the field strength differs from the original value by the amount ( [ deltaf ] ) .
after nucleation , the radius of the bubble grows with constant proper acceleration , and the volume occupied by the new phase keeps increasing .
further nucleation events take place in the region with a low @xmath1 , lowering @xmath0 even further . in the absence of gravity , this
neutralization " would proceed as long as @xmath34 , wiping out any large initial value of @xmath3 and leaving us with a large negative cosmological constant @xmath35 .
of course , this is not what we want .
gravitational effects can improve the picture dramatically @xcite .
in particular , tunneling from flat or anti - de sitter space ( with a negative cosmological constant ) is forbidden provided that the squared tension of the branes is sufficiently large compared with the jump @xmath36 in energy density across the brane , @xmath37 here @xmath38 is the brane tension and @xmath39 is the square of the reduced planck mass . this gravitational shutdown " of brane nucleation could be useful , since an initially large @xmath0 may eventually get neutralized " to a value which is much smaller than @xmath4 in absolute value @xcite .
suppose that the true cosmological constant is negative @xmath40 .
as long as @xmath41 , branes keep nucleating . but once a vacuum with @xmath42 is reached , the process stops provided that eq .
( [ stability ] ) is satisfied .
after that , the vacuum becomes absolutely stable .
brown and teitleboim conjectured that we may live in one such vacuum , where the effective cosmological constant is expected to be of the order of the energy density gap between neighboring vacua @xmath43 in this vacuum , @xmath44 , and the huge negative bare cosmological constant is almost completely cancelled by the @xmath3 contribution in the final state . for sufficiently small charge @xmath33 , eq .
( [ geo ] ) leads to a suppression of @xmath45 relative to the true vacuum energy @xmath46 , which might be helpful in solving the `` old '' fine tuning problem of the cosmological constant .
particle physics models suggest @xmath47 .
hence , we need @xmath48 so that the final value @xmath45 is consistent with the observed value @xmath49 . the constraint ( [ smallcharge ] ) seems rather demanding , since in the context of supergravity we would expect @xmath33 to be closer to the planck scale .
this is the so - called gap problem " .
fmsw argued that the smallness of the charge could be due to the wrapping of branes on degenerating cycles in the extra dimensions .
a successful implementation of this idea has not yet been presented , but some plausibility arguments have been given in @xcite .
alternatively , in a different context , it has been suggested that branes with a very tiny charge @xmath33 may arise due to symmetries of the theory .
an explicit example was given in @xcite , where the branes are not fundamental objects but domain walls of a broken discrete symmetry .
this same symmetry suppresses the coupling of the domain walls to the four - form @xmath1 without any fine - tuning of parameters ( see also @xcite for a fuller discussion ) .
a more severe problem of the brown and teitelboim neutralization scenario , is the `` empty universe '' problem which we discussed in the introduction . by combining the condition ( [ smallcharge ] ) with the stability condition ( [ stability ] )
, it can be shown that the time required to reach the value ( [ geo ] ) is huge compared with the age of the universe @xcite . by the time the effective cosmological constant would be wiped out , all other forms of matter would have also been diluted exponentially , in clear contradiction with observations . furthermore , the endpoint of neutralization would be a state with vanishing or negative effective cosmological constant , whereas the observed value @xmath50 is positive .
one way around the empty universe problem is to consider a slightly different scenario , where the effective cosmological constant is randomized " ( rather than neutralized ) during inflation .
assume , for simplicity , that the energy scale of inflation is much larger than @xmath51 .
in an inflationary phase , brane nucleation processes may increase as well as decrease the value of @xmath3 @xcite .
thus , @xmath0 will randomly fluctuate up and down the ladder as a result of brane nucleation .
inflation is generically eternal to the future and there is an unlimited amount of time available for the randomization process to take place before thermalization @xcite .
assume that the tunneling barriers are sufficiently high , so that no nucleation events happen in the last 60 e - foldings of inflation , or during the hot phase after thermalization up to the present time .
in this scenario there is no empty universe problem : the local value of @xmath0 is decided many e - foldings before the end of inflation , and a wide range of values of @xmath0 will be found in distant regions of the universe , separated from each other by distances much larger than the present hubble radius . some of these regions will just happen to have a very tiny @xmath0 . in combination with anthropic
selection effects , this approach may be used to explain the smallness of the observed effective cosmological constant @xcite .
this may also explain the so - called cosmic time coincidence , or why do we happen to live at the time when an effective cosmological constant starts dominating @xcite .
bousso and polchinski @xcite have proposed a somewhat related randomization " scenario which , moreover , does not rely on branes with the exceedingly small charge @xmath33 satisfying ( [ smallcharge ] ) . in the context of string theory one may expect not just one but many different four - form fluxes @xmath52 coupled to branes with different charges @xmath53 .
each one of these fluxes is quantized in units of the charge , so that @xmath54 . in this case
, the condition for a generic negative cosmological constant to be compensated for by the fluxes is @xmath55 .
the larger the number of different fluxes , the denser is the discretuum of possible values of @xmath0 , and the easier it is to find a set of values of @xmath56 such that the above inequality is satisfied .
a sufficiently dense discretuum is typically obtained provided that the number of fluxes is sufficiently large @xmath57 , even if the individual charges @xmath53 are planckian ( a smaller number of fluxes @xmath58 may be enough in a scenario with large extra dimensions , where the charges @xmath53 are suppressed with respect to the planck scale by a large internal volume effect ) . in the bousso - polchinski model ,
@xmath0 is typically very large , and drives an exponential expansion at a very high energy scale .
suppose we start with a single exponentially expanding domain characterized by a set of integers @xmath59 .
whenever a brane of type @xmath60 nucleates in this region , the integer @xmath61 will change by one unit inside of the brane .
the newly formed region will itself expand exponentially , creating a huge new domain .
the nucleation of further branes within this region will cause an endless random walk of the values of @xmath59 , which will sample the whole discretuum of values of @xmath0 .
eventually , a bubble will nucleate where @xmath0 is comparable to the observed value .
the nucleation of this last bubble is still a high energy process , which kicks some inflaton field off its minimum , and starts a short period of inflation within this last bubble .
this period of ordinary " inflation is necessary in order to produce the entropy we observe , thereby avoiding the empty universe problem .
of course , some anthropic input is still necessary in this approach in order to explain why , out of the discretuum of possibilities , we live in a vacuum with small @xmath0 .
nucleation of coincident branes would drastically modify the neutralization scenario of brown and teitelboim , as well as the randomization scenarios sketched above . as proposed in @xcite , in the case of neutralization
, this modification may lead to a solution of the empty universe problem . in the randomization scenarios
there is no such problem , and it is unclear whether an enhancement of the multiple brane nucleation rate is desirable at all .
this enhancement would trigger large jumps in the effective cosmological constant , making the calculation of its spatial distribution more complicated than for single brane nucleation .
thus , it is of interest to understand the conditions under which multiple brane nucleation is allowed , and what are the degeneracy factors which might enhance their nucleation rate relative to the nucleation of single branes . before addressing this issue
, it will be convenient to present a short discussion of coincident branes .
in the context of string theory , one may consider stacks of @xmath7 coincident d - p - branes .
each brane in the stack has charge @xmath33 with respect to the form field @xmath30 .
thus , in four noncompact dimensions , a pair of parallel 2-branes repel each other with a constant force per unit area given by @xmath62 due to the four - form field interaction .
in the ten dimensional theory , the repulsive force due to @xmath63 is balanced with other contributions from the closed string sector , such as the graviton and dilaton . as a result , there are no net forces amongst the different branes on the stack .
it is in principle possible to maintain this delicate balance by suitable compactification from 10 to 4 dimensions . in 4d ,
the branes look like domain walls , and their interaction through the ordinary graviton leads to a mutual repulsive force given by @xcite , @xmath64 hence , both forces given by ( [ qfield ] ) and ( [ dwfield ] ) tend to push the branes apart from each other . on the other hand , some of the higher dimensional gravitational degrees of freedom are represented by scalars in 4d , and these , together with the dilaton , lead to attractive forces .
let us first consider the case of d-2-branes in 10 dimensional type iia supergravity .
the relevant part of the action is given by ( see e.g. @xcite ) @xmath65 - t_2\int_\sigma d^3\xi \sqrt{g_\sigma}\ e^{-\phi } + \hat q_2 \int_\sigma d^3\xi\ \hat{\cal a}.\ ] ] here , @xmath66 and @xmath67 are the determinants of the 10 dimensional metric @xmath68 and of the metric induced on the worldsheet @xmath69 , respectively , whereas @xmath13 is the ricci scalar corresponding to @xmath68 .
the hats on the four - form , the gauge potential , and the corresponding charge , are introduced in order to distinguish them from the four dimensional ones which will be used below , and which differ from those by constant normalization factors .
compactifying on a calabi - yau manifold @xmath70 through the ansatz @xmath71 where the greek indices run from 0 to 3 , we readily obtain the following four - dimensional action : @xmath72 @xmath73 here , @xmath74 , where @xmath75 is the coordinate volume of the manifold @xmath70 , @xmath76 is the ricci scalar for the metric @xmath25 , and we have introduced two linear combinations of the internal volume modulus @xmath77 and the dilaton @xmath78 @xmath79 the field @xmath80 decouples from the branes , and shall be ignored in what follows . in ref.@xcite ,
a different expression was given for the dimensionally reduced action , because no modulus was introduced for the size of the internal space .
however , as we shall see , such modulus is necessary for the cancellation of forces among the branes . a more direct route to the 4-dimensional theory ( [ reduced4d ] ) , which will be slightly more convenient for our discussion , is to start directly from 11-dimensional supergravity and compactifying on a 7 dimensional internal space .
the action in eleven dimensions is given by @xmath81 - t_2\int_\sigma d^3\xi \sqrt{g_\sigma}+ \hat q_2 \int_\sigma \
\hat{\cal a}.\ ] ] where @xmath82 . introducing the ansatz @xmath83 with a ricci flat internal manifold ,
we find @xmath84 where @xmath85 and @xmath86 not surprisingly , this has the same form as ( [ reduced4d ] ) , since after all the 10-dimensional lagrangian can be obtained from the 11-dimensional one by compactifying on a circle .
( [ reduced4d2 ] ) is a particular case of the slightly more general action in four dimensions : @xmath87 - { 1\over 4!}\ e^{-2\alpha{\varphi } } { \cal f}^2 \right\}- \sigma \int_{\sigma } d^3 \xi \sqrt{\gamma}\ e^{\alpha { \varphi } } + q \int_{\sigma } { \cal a } .
\label{action4d}\ ] ] the parameter @xmath88 characterizes the scalar charge of the brane . as we shall see below , linearizing in @xmath89 around @xmath90 , it can be easily shown that the scalar force is given by @xmath91 where we have introduced the scalar charge @xmath92 thus , from ( [ qfield ] ) , ( [ dwfield ] ) and ( [ scalarforce ] ) , the branes will be in indifferent equilibrium provided that the following relation holds : @xmath93= 0.\label{q2}\ ] ] from ( [ alpha ] ) , this condition can be rewritten as @xmath94 note that the case of type iia supergravity discussed above , corresponds to @xmath95 , @xmath96 , and @xmath97 [ note that @xmath98 .
therefore ( [ bog ] ) is satisfied provided that @xmath99 , the usual bps condition . multiple brane solutions to ( [ action4d ] )
can easily be constructed provided that ( [ bog ] ) is satisfied . in the bulk , the equation of motion for @xmath30 leads to @xmath100 where @xmath1 is a constant and @xmath26 is the levi - civita symbol . at the branes , this constant jumps by the amount
@xmath101 it can be checked that the remaining equations of motion for the scalar and the gravitational field follow from the action @xmath102 - v_i(\varphi ) \right\ } - \sigma \int_{\sigma } d^3 \xi \sqrt{\gamma}\ e^{\alpha { \varphi } } , \label{action4dii}\ ] ] where @xmath103 and the sum is over the regions with different values of @xmath1 . with the metric ansatz @xmath104 where latin indices run from 0 to 2 , the solution is given by @xmath105 and @xmath106 here , @xmath107 are integration constants .
these , and the sign option in ( [ branesolution ] ) , must be chosen so that the junction conditions for the gravitational field and for the scalar field are satisfied at the branes .
for the gravity part , the condition is @xcite , @xmath108 = -4\pi g\sigma\ e^{\alpha { \varphi } } \gamma _ { ab } , \label{discont}\ ] ] where @xmath109 $ ] is the difference of extrinsic curvature on the two sides of the brane and @xmath110 is the world - sheet metric . in the present case
, this reduces to @xmath111 = -{\sigma\over 2 m^2_p}\ e^{\alpha { \varphi}}. \label{junctiona}\ ] ] for the scalar field , the junction condition which follows from the equation of motion at the brane reduces to @xmath112= \alpha{\sigma\over m_p^2}\ e^{\alpha { \varphi } } , \label{junctionphi}\ ] ] which is consistent with ( [ junctiona ] ) and ( [ solutiona ] ) .
for instance , a solution with two branes separated by a distance @xmath113 is given by flat space ( @xmath114 ) in the region between the branes , and by @xmath115 in the exterior region ( see fig [ fibranas ] ) .
the solution corresponds to two branes of charge @xmath33 interpolating between regions with @xmath116 and @xmath117 , separated by a region with @xmath118 .
[ ] [ r]@xmath116 [ ] [ l]@xmath117 [ ] [ ] @xmath118 note that the solution ( [ sandwich ] ) contains a singularity at a distance @xmath119 ( where the warp factor @xmath120 and the volume of the internal space vanish ) . in what follows , however , we shall not be interested in flat , infinite branes , such as the ones discussed above .
rather , we shall be interested in compact instanton solutions of finite size , in a theory with a stabilized dilaton . in this case , the singularities due to linear potentials of the form ( [ sandwich ] ) should not arise , but the solutions discussed above should remain a good description in the vicinity of the branes .
another consequence of stabilizing the dilaton is that the perfect balance of forces amongst the branes will be spoiled at distances larger than the inverse mass of the moduli , as we now discuss .
as it stands , the dimensionally reduced supergravity lagrangian ( [ reduced4d ] ) [ or its generalization ( [ action4d ] ) ] is not useful for discussing the neutralization of the cosmological constant .
the lagrangian does not include the bare cosmological term @xmath4 , which is precisely the subject of our interest , and the term proportional to @xmath3 does not behave as an effective cosmological constant , but rather as an exponential potential for a modulus @xmath89 ( which is not flat enough to mimick the vacuum energy ) have been thoroughly studied in the literature , and it is known that they can drive cosmological solutions with a power - law scale factor @xcite .
such attractor solutions are approached for a wide range of initial conditions , and the resulting expansion can be accelerating or decelerating , depending on whether @xmath121 or @xmath122 . for @xmath123
the cosmological scale factor approaches @xmath124 , where @xmath125 is cosmological time .
this solution corresponds to an effective equation of state @xmath126\rho$ ] , where the ratio of kinetic and potential energies of the scalar field @xmath89 remains constant . in our case , from ( [ bog ] ) , we need @xmath127 and therefore the kinetic term becomes completely dominant in the long run , which leads to @xmath128 .
hence , by itself , the @xmath3 term does not behave like an effective cosmological constant . ] ) .
a more realistic model is obtained by introducing a stabilization mechanism which fixes the expectation value of @xmath89 , and gives it some mass @xmath129 .
once @xmath89 is stabilized the @xmath3 term does behave as a contribution to @xmath0 .
stabilization is also desirable because the dilaton and the radion moduli ( corresponding to the size of extra dimensions ) mediate scalar interactions of gravitational strength , which are severely constrained by observations .
the study of mechanisms for stabilization is currently an active topic of research ( see e.g. @xcite and references therein ) . although the details of stabilization will not be too important in our subsequent discussion , it may be nevertheless illustrative to have in mind a specific toy mechanism ( for which we do not claim any rigorous justification in the context of string theory ) .
the general problem is that the potential ( [ run ] ) has no minimum and leads to a runaway dilaton . in order to create a minimum ,
let us consider two contributions which may be added to ( [ run ] ) .
first of all , instead of using a ricci flat internal manifold in ( [ ansatz11 ] ) we may compactify on an einstein manifold , with @xmath130 upon dimensional reduction , the curvature @xmath70 contributes an exponential term to the effective potential for @xmath89 .
however , this will still not be sufficient for stabilizing the internal volume in an interesting way .
in fact , as shown by nunez and maldacena @xcite , there are no static compactifications of the classical supergravity lagrangian with a positive effective four - dimensional cosmological constant @xmath131 , and so far we have added nothing to the classical lagrangian .
thus , in order to implement the four dimensional situation of our interest , a third term related to quantum corrections has to be considered .
following candelas and weinberg @xcite , we may consider the casimir energy of bulk fields ( we are of course assuming that supersymmetry is broken , so that the casimir contributions of bosons and fermions do not cancel each other exactly ) . in the example considered by candelas and weinberg @xcite , besides the casimir energy term , a higher dimensional cosmological term @xmath132 was used , and the internal manifold was taken to be a space of constant positive curvature ( @xmath133 ) . in 11-d supergravity a cosmological constant is not allowed , so here we use the @xmath3 flux instead . also , we will need to compactify on a negatively curved internal manifold ( @xmath134 ) , or else the dilaton
would be stabilized at negative @xmath0 .
the versatility of negatively curved compactifications has been stressed in @xcite .
in particular , they have the interesting property of rigidity , which means they do not lead to other moduli besides the size of the internal space .
compact hyperbolic manifolds ( chm ) can be obtained from the maximally symmetric negatively curved space @xmath135 through identifications by a discrete isometry group @xmath136 .
the volume of @xmath137 is given by @xmath138 where @xmath139 is the curvature radius of the manifold , related to the curvature parameter in ( [ einstein ] ) by @xmath140 the factor @xmath141 depends on the topology , and it is bounded below but not above .
if @xmath142 is the largest distance around the manifold , then for @xmath143 we have @xmath144 .
the kaluza klein ( kk ) spectrum in this manifold is believed to have a mass gap , bounded below by @xmath145 from the 11 dimensional point of view the casimir energy density scales like @xmath146 the factor @xmath147 can be estimated by naive dimensional analysis as @xmath148 where @xmath149 is the number of physical polarizations of bulk fields , and @xmath150 is some small one - loop factor .
this factor will depend on the precise topology of the compact hyperbolic manifold , but it could plausibly be in the range @xmath151 . hence , the parameter @xmath147 could be of order one .
explicit calculations for different choices of the manifold @xmath137 have not been performed , and are well beyond the scope of the present paper ( see e.g. @xcite and references therein ) . in what follows we shall leave @xmath147 unspecified , assuming that a compactification exists where @xmath152 . multiplying the higher dimensional energy density @xmath153 given in ( [ rho11 ] ) by the size of the internal manifold , @xmath154 , and by a factor @xmath155 which arises from the four - dimensional volume element in going to the einstein frame , the effective potential which appears in eq .
( [ action4dii ] ) gets replaced by @xmath156 where we have also added the curvature contribution . here
@xmath157 , where @xmath158 .
we can always adopt the convention that @xmath159 , since a change of @xmath160 can be reabsorbed by a shift in @xmath161 and a constant re - scaling of the four - dimensional metric @xmath162 ( in this frame , the curvature of the manifold is of the order of the higher dimensional planck scale for @xmath163 ) . since @xmath164 , where @xmath56 is an integer and @xmath165
, we have @xmath166.\label{stabpot2}\ ] ] as illustrated by eq .
( [ v7 ] ) a large value of @xmath167 in units of @xmath159 can be obtained by using a manifold with sufficiently complicated topology . since @xmath168 , the factor @xmath169 in the last term of ( [ stabpot2 ] ) can be rather small .
the scale @xmath170 could be as low as the @xmath171 , in which case that factor can be as low as @xmath172 .
moreover , as emphasized in @xcite , this can be achieved for chm even if the linear size @xmath142 of the internal manifold is not very much larger than @xmath139 . in fig .
2 we plot the effective potential for @xmath173 , @xmath174 and values of @xmath175 and @xmath176 . when a single brane nucleates , it changes @xmath56 by one unit , and hence changes the value of the effective potential at the minimum ( changing therefore the effective cosmological constant ) .
if the discretuum of values of @xmath52 were sufficiently dense , then there would always be one of the minima of the effective potential where the vacuum energy is sufficiently small to match observations . in the case
we have considered here , the discretuum is not dense enough .
the cancelation between the last term in ( [ stabpot ] ) and the other two requires @xmath177 , and so the gap between levels near @xmath178 can be estimated as @xmath179 , which is far too large .
this situation can be remedied by considering 5 m branes wrapped around 3-cycles in the internal space . as emphasized by bousso and polchinski @xcite ,
these are coupled ( with different charges ) to additional fluxes , and a large number of fluxes will result in a much denser spectrum .
alternatively , fmsw have suggested that the branes may wrap a degenerating cycle @xcite , in which case the individual charges might themselves be exponentially smaller than the fundamental scale , resulting also in a sufficiently dense discretuum . before closing this section
, we should note that the location where the modulus sits is basically determined by the competition between the two first terms in ( [ stabpot ] ) , corresponding to curvature and casimir energy . the physical curvature radius of the internal space is therefore stabilized at @xmath180 , where @xmath181 is the spacetime dimension .
thus , unless the constant @xmath147 in eq .
( [ rho11 ] ) is exceedingly large , the compactification scale is comparable to the inverse of the higher dimensional cut - off scale @xmath170 . in such case the semiclassical analysis
which we have entertained above is not justified , since higher order corrections will be just as important as the one loop effect which we have included .
this appears to be a generic problem when we try to stabilize by making the curvature and the casimir terms comparable , as in the candelas and weinberg example @xcite .
there , the problem was not quite as poignant , since the constant @xmath147 could be made very large by adding a sufficient number of fields ( also , as it is clear from the above estimate of @xmath182 , the problem is somewhat milder if the number of extra dimensions is smaller ) .
here we shall not dwell on this problem , since the main purpose of the above discussion is just to illustrate the role of the four - form in obtaining a discretuum of states . for more fundamental approaches to moduli stabilization in the present context
, the reader is referred to @xcite and references therein .
our ensuing discussion will be largely independent of the details of the stabilization mechanism .
inevitably , the stabilization of @xmath89 spoils the perfect balance of forces . at distances larger than the inverse of the mass of the modulus @xmath183 , the remaining gravitational and four - form interactions are both repulsive , and lead to a linear potential per unit surface of the form @xmath184\ d \quad\quad \quad ( d\gg m^{-1})\ ] ] where @xmath113 is the distance between the branes . to investigate the behaviour of the interaction potential at distances shorter than @xmath183 ,
let us first consider the situation where gravity and the 3-form gauge potential @xmath30 are ignored , and the branes interact only through a scalar field @xmath89 of mass @xmath129 .
the action is given by : @xmath185 % + q \int_{\sigma } { \cal a } -\sigma \int_{\sigma } d^3 \xi \sqrt{\gamma}\ e^{\alpha\varphi}. \label{actiontoy3}\ ] ] the solution with a single brane on the plane @xmath186 has the cusp profile @xmath187 where @xmath188 is a solution of @xmath189 the energy per unit area of this configuration is given by @xmath190 for small charge and tension , we have @xmath191 due to the scalar field dressing , the effective tension of the brane , denoted by @xmath192 , is smaller than the parameter @xmath38 which appears in the action .
this effect becomes more dramatic if we place a large number @xmath7 of branes on top of each other . since both @xmath38 and @xmath193 scale like @xmath7 , the effective tension of the stack is given by @xmath194,\quad\quad ( k\ e^2 \ll 2 m \sigma ) , \label{sigmak}\ ] ] which grows with @xmath7 but less than linearly . in the limit of very large @xmath7 , using ( [ transcendental ] ) with @xmath38 replaced by @xmath195 we have @xmath196 and so the tension almost saturates , growing only logarithmically with @xmath7 .
this last expression should not be taken too literally , however , since in ( [ actiontoy3 ] ) we are neglecting gravitational effects .
as shown above , nonlinear effects of gravity become important at a distance given by @xmath197 , which in the limit given by ( [ logarithm ] ) is smaller than @xmath183 . hence the cusp solution ( [ cusp ] ) , which has typical width @xmath198 , will receive sizable gravitational corrections in the limit of large @xmath7 .
nevertheless , it still seems likely that the effective tension of the stack of branes will grow with @xmath7 much slower than linearly .
physically , the reason is that the scalar charge of the stack increases with @xmath7 . according to ( [ junctionphi ] ) , this means that the cusp in the field @xmath89 on the branes grows stronger , which in turn means that the value of the field @xmath188 on the branes has to be further displaced into negative values .
hence , the brane contribution to the effective tension @xmath199 shows only a very modest growth with @xmath7 , and the tension for large @xmath7 is in fact dominated by the potential and gradient energy of the scalar near the brane .
let us now look at the interaction potential between two branes separated from each other . for simplicity , we shall restrict attention to the case of small scalar charge , @xmath200 .
then , the last term in ( [ actiontoy3 ] ) is well approximated by its linearized expression : @xmath185 % + q \int_{\sigma } { \cal a } - \int_{\sigma } d^3 \xi \sqrt{\gamma}\ ( \sigma + e m_p\ \varphi ) .
\label{actiontoy}\ ] ] placing the two branes at @xmath201 , the solution for the scalar field has the golden gate " profile shown in fig .
3 @xmath202 @xmath203 from this expression , and adding the long range contributions from gravity and the four - form , the interaction potential per unit surface is given by @xmath204\ d - { e^2\over 2 m } e^{-md}.\ ] ] at short distances , this takes the form @xmath205 where the parameter @xmath93,\label{q22}\ ] ] was introduced in ( [ q2 ] ) .
as shown in section iii , dimensional reduction from the supergravity lagrangian with bps charges gives @xmath206 . in this case
, the linear term in ( [ interaction2 ] ) disappears .
the quadratic term is negative , which means that the stack of coincident branes is unstable and tends to dissolve .
thus the stabilization of the modulus @xmath89 seems to make the superposition of branes an unstable configuration . as we shall see in more detail in section vii
, stacks of branes in marginal or unstable equilibrium will not be appropriate for constructing instantons , since in particular these would have too many negative modes .
the instanton is only meaningful if the branes attract each other .
the above analysis shows that at the classical level , the branes with bps values of the charges would not attract each other , and consequently the nucleation of coincident branes is not allowed at least in the semiclassical description
. there may be several escape routes to this conclusion .
for instance , after supersymmetry breaking , the charges of the branes get renormalized , and it is possible that the corrected charges satisfy @xmath207 . in this case
the branes in the stack would attract each other with a linear potential .
other mechanisms by which nearby branes attract each other are conceivable , but here we shall not try to pursue their study
. it should be emphasized , however , that this remains an important open question which needs to be addressed in order to justify the semiclassical description of coincident brane nucleation . in section vii , where we discuss the degeneracy factor , we shall simply postulate that an attractive interaction exists at short distances .
the brane nucleation rate per unit volume is given by an expression of the form @xcite @xmath208 where @xmath209 . here
@xmath210 is the euclidean action of the instanton corresponding to the decay of the four - form field , and @xmath211 is the action for the background solution before nucleation . the prefactor @xmath212 will be discussed in the next section .
the formal expression ( [ nucrate ] ) can be used both in flat space and in curved space .
also , it can be used at zero or at finite temperature .
the difference is in the type of instanton and background solutions to be used in each case . in this section
, we shall concentrate in the maximally symmetric instanton . in flat space , this represents the decay of a metastable vacuum at zero temperature .
the euclidean solution can be described as follows @xcite . at infinity ,
the field strength takes the value @xmath213 , which plays the role of a false vacuum . near the origin
, we have @xmath214 .
this plays the role of a true vacuum phase .
both phases are separated by the euclidean worldsheet of the brane , which is a three - sphere of radius @xmath215 here , @xmath38 is the tension of the brane and @xmath216\ ] ] is the jump in the energy density accross the brane .
the difference in euclidean actions between the instanton and the background solution is given in this case by @xcite @xmath217 if instead of considering a single brane , we are looking at the nucleation of @xmath7 coincident branes , an analogous solution of the euclidean equations of motion should be considered .
the only difference is that the effective charge is a factor of @xmath7 higher , and the effective tension is also higher by approximately the same factor . using ( [ sigmak ] )
, we should replace @xmath218,\ ] ] @xmath219 , \label{sk2}\ ] ] in eq .
( [ bflat ] ) for the bounce " action . as shown in section
v , the approximate form of @xmath220 is valid when the second term in the r.h.s . of ( [ sk2 ] )
is small compared with the first . in the model considered in section iv , the modulus is stabilized with a mass of order @xmath221 .
this may be regarded as a conservative upper bound for the general case . since @xmath222 ,
the condition @xmath223 requires @xmath224 for larger @xmath7 , gravitational corrections to the brane profile become important on a lengthscale comparable to the inverse mass of the dilaton , as we discussed in the previous section . in this case
the instanton solution will depend on the detailed dynamics of the dilaton near the brane .
the investigation of these dilatonic instantons is per se an interesting problem , which we leave for further research . here
, for simplicity , we shall assume a scenario where either @xmath225 , due perhaps to the wrapping of branes on a degenerating cycle , or where the extra dimensions are relatively large , so that @xmath226 , and we shall restrict attention to instantons where the number of branes is bounded by ( [ kll ] ) .
this gives @xmath227 ^ 3}. \label{bflatk}\ ] ] nucleation in flat space is impossible when the number of branes is too large , since otherwise we would be jumping to vacua with a higher energy density .
therefore , we must restrict to @xmath228 .
in fact , the minimum value of @xmath52 is achieved for the largest @xmath7 satisfying @xmath229 .
note that the action increases faster than linearly as we increase the number of branes @xmath7 , but we still have @xmath230 throughout this range .
when gravity is taken into account , the maximally symmetric instanton was given by coleman and de luccia @xcite .
it is constructed by gluing two different de sitter solutions ( that is , two four - spheres ) at the worldsheet of the brane , which is still a three - sphere .
the instanton is sketched in fig .
[ coldel ] .
the four - spheres have the radii @xmath231 and @xmath232 , where @xmath233 are the hubble rates of the de sitter phases before and after the nucleation event , respectively . here , and in what follows , we are assuming that the effective cosmological constant is still positive after nucleation , since these are the final states we are interested in .
the bubble radius at the time of nucleation ( which coincides with the radius of the three - sphere ) is bounded by @xmath234 .
analytic expressions for @xmath13 are given in refs.@xcite .
the general expression is cumbersome and not particularly illuminating , so we shall concentrate on a few limiting cases of interest .
[ ] [ r]@xmath235 [ ] [ l]@xmath236 [ ] [ ] @xmath237 as discussed in section iii , the gravitational field of a brane is repulsive , and is characterized by an acceleration " of order @xmath238 .
this gravitational field will be negligible provided that the corresponding rindler radius ( or inverse of the acceleration ) is much larger than the radius @xmath13 of the euclidean worldsheet , which in turn is smaller than @xmath231 @xmath239 in this regime , we can distinguish two cases . for @xmath240 ,
the radius of the euclidean worldsheet is much smaller than the de sitter radius , and the flat space expression ( [ bflat ] ) holds . in the opposite limit , @xmath241
, we have @xmath242 and @xmath243 the vacuum energy difference @xmath26 is unimportant in this case , and the action coincides with that for domain wall nucleation @xcite .
finally , the gravitational field of the brane is important when @xmath244 in this case , the radius of the worldsheet is given by @xmath245 , and @xmath246 in this limit , the action of the instanton is much smaller than the action of the background , and this is the reason why ( [ bwallgrav ] ) is independent of the tension .
the same arguments apply of course to the instantons with coincident branes , and the corresponding expressions for the action and radii in the different regimes are summarized in table i. @xmath247 + @xmath248 & .values of the radius @xmath13 of the coleman - de luccia instanton and the corresponding bounce action @xmath249 in different limits [ cols= " < " , ] & @xmath250 +
the prefactor @xmath212 in eq .
( [ nucrate ] ) is given by @xcite @xmath251 here , @xmath252 and @xmath253 are the gaussian integrals of small fluctuations around the instanton and the background solutions respectively . expanding all brane and bulk fields ( which we generically denote by @xmath78 ) around the instanton configuration @xmath254 as @xmath255 , we have @xmath256= s_e[i ] + s^{(2)}[i,\delta\phi_j ] + ... , $ ] where @xmath257 includes the terms quadratic in @xmath258 . at the one loop order we have @xmath259 in the functional integral , there are some directions which correspond to spacetime translations of the instanton .
the primes in the numerator of ( [ pref ] ) and in the previous equation indicate that the translational zero modes are excluded from the integration , and replaced by the corresponding spacetime volume .
the latter is subsequently factored out in order to obtain a nucleation rate per unit time and volume .
the degrees of freedom which live on the brane will make a contribution to the numerator but not to the denominator .
consider , for instance , a free bosonic field @xmath260 of mass @xmath261 living on the worldsheet .
its contribution to the prefactor is @xmath262 here , the integral in the exponent is over the euclidean worldsheet of the brane .
if we have @xmath263 of such fields , their effect on the nucleation rate is to replace @xmath264 in the naive expression for the nucleation rate , @xmath265 , a replacement which can become very important as we increase the number of fields .
this is , in essence , the observation made by feng et al . that the large number of worldsheet fields might strongly affect the nucleation rate .
the euclidean worldsheet in the coleman - de luccia instanton is a 3-sphere .
determinantal prefactors due to scalar fields were considered in some detail in @xcite .
they are given by @xmath266 where @xmath267 and @xmath268 is the usual riemann zeta function .
for instance , the contribution of a conformally coupled scalar field can be obtained by taking @xmath269 , which gives @xmath270 hence , the effective degeneracy factor contributed by a conformal scalar field is given by @xmath271 the first thing to note is that this factor is not an enhancement , but a suppression .
hence , the prefactor can not simply be thought of as the exponential of an entropy . more generally , from eq .
( [ zetadegen ] ) , the prefactor is independent of the expansion rate of the ambient de sitter space .
this has implications for the scenario proposed by fmsw , as we shall discuss in the next section .
in general , the degeneracy factor will depend on @xmath13 and on the mass of the particle . for light minimally coupled scalars , equation ( [ zetadegen ] )
gives @xmath272 there can be a strong enhancement in the nucleation rate if there are very light scalar fields . in the limit
@xmath273 the factor goes to infinity .
this is because a massless scalar has a normalizable zero mode on the sphere , corresponding to the symmetry @xmath274 . in this case , the zero mode must be treated as a collective coordinate . the nucleation rate is proportional to the range @xmath275 of the field @xmath260 , because the bubbles can nucleate with any average value of the scalar field with equal probability @xcite @xmath276 ( \pi r^3)^{1/2 } \delta \phi = e^{{\zeta_r}'(-2 ) } r^{1/2 } \delta \phi .
\label{dsmassless}\ ] ] finally , for large mass the expression ( [ zetadegen ] ) leads to @xmath277 the exponent of this expression can be interpreted as a renormalization of the tension of the stack of branes , due to the heavy scalars living on it @xcite .
indeed , the effective potential for a scalar field in 2 + 1 dimensions in the flat space limit is proportional to @xmath278 , and the factor of @xmath279 is just due to the volume of the worldsheet .
the factor @xmath280 is plotted in fig . [ prefactor ] for different values of @xmath281 . [ ] [ ] @xmath280[][]@xmath282 note that there is an enhancement both at smal and at large mass , but the two have very different origin .
the large value of ( [ inv ] ) for light fields can be interpreted as a phase space enhancement . as we shall discuss in subsection vii - d , quantum fluctuations of fields living on the worldsheet of the brane
are characterized by a temperature @xmath283 .
the corresponding fluctuations in the potential term are of order @xmath284 , which corresponds to a root mean squared expectation value for @xmath260 of the order @xmath285 eq .
( [ inv ] ) is recovered if we insert the range ( [ delphik ] ) in eq .
( [ dsmassless ] ) .
the lighter the field , the larger is the phase space factor @xmath275 , and the larger is the nucleation rate .
for @xmath286 this argument can not be used , since the field does not behave as effectively massless . in
that limit the field decouples , as it should , and its effect is felt as a renormalization of the parameters in the lagrangian . for a scalar field ,
the leading effect is to renormalize the tension of the brane , making it lower ( see appendix a ) .
this causes an exponential enhancement of the nucleation rate . besides scalar fields
, gauge bosons and fermions may live on a brane . at low energies ,
the field content on a stack of coincident branes will be model dependent .
the idea is that it will correspond to a gauge theory whose symmetry is enhanced when branes are coincident , giving rise to a large number of light species on the worldsheet .
the details of the theory , however , will depend on whether we start from 2d branes which descend directly from the ten dimensions , of from higher dimensional p - branes wraped on ( p-2)-cycles .
they will also depend on the details of compactification .
rather than building a particular scenario from first principles , here we shall try to gain some intuition by considering a toy model directly in four dimensions .
the degrees of freedom on the stack of @xmath7 coindident branes will be bluntly modeled by a weakly coupled susy @xmath8 gauge theory on the 2 + 1 dimensional worldsheet .
this contains a @xmath8 gauge field , @xmath287 scalar degrees of freedom in the adjoint representation of @xmath288 , and a scalar singlet ; plus the corresponding fermionic degrees of freedom .
the action for the worldsheet fields is given by @xmath289 , \label{ym}\ ] ] where the ellipsis indicate the terms containing fermions . here
@xmath290 is the field strength of the gauge field @xmath291 , where @xmath292 are the generators of @xmath8 , normalized by @xmath293 , and @xmath294 , where @xmath295 are real scalar fields @xmath296 . by analogy with the well known case of d - branes in 10 dimensions
, we shall assume that if the coincident branes are flat ( as in the case when there is no external four - form field ) , then the theory is supersymmetric and all degrees of freedom are massless .
the scalar field @xmath78 is a hermitian matrix and can always be diagonalized by a suitable @xmath8 gauge transformation , @xmath297 .
the eigenvalues @xmath298 , @xmath299 are then interpreted as the positions of the different parallel branes along an axis perpendicular to them ( if the codimension of the branes were higher , there would be additional scalar matrices representing the positions of the branes along the additional orthogonal directions , but here we are interested in the case of codimension one ) . in the supersymmetric case ,
the potential for the scalar field vanishes , @xmath300 however , for non - flat branes , the displacement of the stack of branes no longer behaves as an exatly massless field @xcite , but one which couples to a combination of the worldsheet and extrinsic curvatures , as well as to the background four - form field .
also , after moduli stabilization , the forces amongst branes are nonvanishing , and this also contributes to the potential for the relative displacements of the different branes .
thus , as we shall see , there will be a nonvanishing potential @xmath301 in the physical situation of our interest .
when the positions are not coincident , the @xmath8 symmetry breaks to a smaller group because some of the gauge bosons aquire masses @xmath302 .
there are always at least @xmath7 massless vectors ( corresponding to maxwell theory on the individual branes on the stack ) and the remaining @xmath303 have double degeneracy . for example , in the case of a single brane , the scalar field will represent the goldstone mode of the broken translational symmetry , associated to transverse displacement of the brane . for the case of two branes , there are two such scalars
. one of them , @xmath304 , will corresponds to simultaneous motion of both branes , and is a singlet under @xmath305 .
the other one , @xmath306 , will correspond to relative motion of the branes and it transforms under @xmath305 .
when the two branes move apart , @xmath307 acquires an expectation value and two of the gauge bosons get a mass , breaking the symmetry @xmath308 .
the case of interest to us is not a flat brane , but the world - wolume of the @xmath309-brane in the cdl instanton , which forms a 3-sphere of radius @xmath13 . in this situation
, we do not expect the theory to be supersymmetric ( in particular , corrections to the effective action will appear at one loop , which will be related in fact to the determinantal prefactor in the nucleation rate ) .
the case of a single brane is very similar to the case of a vacuum bubble , and in that case we know that the transverse displacements correspond to a scalar field of negative mass squared @xcite @xmath310 the origin and precise value of this mass term can be understood geometrically , since it leads to four normalizable zero modes which are the spherical harmonics with @xmath311 .
these correspond to the four space - time translational modes of the instanton , which have to be treated as collective coordinates .
this scalar field has also a single negative mode , which is the constant @xmath312 mode .
a negative mode is precisely what is needed for an instanton to contribute to the imaginary part of the vacuum energy , and hence to contribute to false vacuum decay @xcite .
integrating out the transverse displacement of the brane gives a determinantal prefactor of the form @xcite @xmath313 where @xmath314 is the spacetime volume .
the prefactor in the nucleation rate ( [ nucrate ] ) per unit time and volume is obtained after dividing by @xmath315 .
the above argument neglects gravity , and it is a good approximation when @xmath316 . the case of strong gravity @xmath317 is far more complicated , since one has to integrate over fluctuations of the gravitational field in the bulk , and is left for further research .
if there are 2 coincident branes , then there are two independent transverse displacements corresponding to the eigenvalues @xmath318 and @xmath319 .
center of mass " displacement @xmath320 behaves just like in the case of a single brane .
the orthogonal combination @xmath321 represents the brane separation , and the two - brane instanton will only be relevant if this second combination acquires a positive mass through some mechanism , so that there is a single negative mode , not two , and four normalizable zero modes in total . in other words , the branes must attract each other . as shown in section v ,
if the branes have bps charges , they in fact tend to repel each other once the dilaton is stabilized , and the configuration with coincident branes is unstable . in this case
, we do not expect that there will be any instanton representing the nucleation of multiple branes .
one possible way out of this conclusion , is to assume that the charges are different from their bps values , due to supersymmetry breaking effects , so that the sum @xmath322 defined in eq .
( [ q22 ] ) is positive . in that case , the two branes attract each other with a linear potential .
the canonical field @xmath321 is related to the interbrane distance @xmath113 by @xmath323 .
hence , the interbrane potential ( [ interaction2 ] ) takes the form @xmath324 this potential is attractive at small distances , and has a maximum at @xmath325 .
classically , the branes will attract at short distances .
however , there is a danger that they will be separated by quantum fluctuations . as we discussed in the previous subsection ( and
we will argue more at length in subsection d ) quantum fluctuations of fields living on the worldsheet of the brane are characterized by a temperature @xmath283 .
the fluctuations in the potential are of order @xmath326 , and these correspond to a root mean squared expectation value for @xmath321 of the order @xmath327 the stability of the two - brane instanton requires that @xmath328 .
otherwise , unsuppressed quantum fluctuations take the field over the barrier and the distance between the branes starts growing without bound .
this requires @xmath329 if this condition is satisfied , then @xmath321 is trapped near the origin , and the branes stay together .
it can be shown that for @xmath330 the field behaves as approximately massless in the range given by ( [ deltaphim ] ) , so from ( [ dsmassless ] ) its contribution to the prefactor can be estimated as @xmath331 as in the case of the massive field discussed in the previous section , this expression is only justified when @xmath332 .
if @xmath322 is too large , then the field will not behave as massless in the range ( [ deltaphim ] ) , and we expect that for @xmath330 the sole effect of the field will be to renormalize the coefficients of operators such as the brane tension in the classical lagrangian
. an inconvenient feature of the linear potential ( [ interaction3 ] ) is that is non - analytic at the origin , and hence an explicit calculation in the limit of large slope is not straightforward .
moreover , we can not write down an expression for it in terms of the matrix operator @xmath78 , but just in terms of its eigenvalues @xmath298 . another possibility , which is more tractable from the formal point of view , is to assume that there is an attractive interbrane potential which is quadratic at short distances . that is , as in ( [ interaction3 ] ) with @xmath206 but with a positive coefficient in front of the second term . in terms of the eigenvalues @xmath298 , which represent the displacements of the branes , we assume the following expression for the potential @xmath333 where @xmath334 and @xmath335 [ as given in eq .
( [ nmass ] ) ] , while @xmath336 is a new parameter which characterizes the attractive interaction at short distances . in terms of the field
@xmath78 , we can write the potential as @xmath337\ ] ] @xmath338 where in the last equality we have expanded @xmath294 in the basis of generators @xmath292 , and we have used @xmath339 and @xmath340 for @xmath341 . in the symmetric phase ( and assuming @xmath342 ) each one of the adjoint fields @xmath343 will contribute a determinantal prefactor of the form @xmath344 where we have used ( [ inv ] ) .
this shows a somewhat milder dependence in @xmath13 than in the case of a linear interaction between branes , given in eq .
( [ dml ] ) , but still of power law form . in the limit of large mass @xmath345 ,
which as discussed before amounts to a finite renormalization of the brane tension . aside from scalars
, we should also consider the contributions from gauge bosons and fermions .
for the case of a 3-sphere , these have been studied in @xcite . for vectors of mass @xmath346 ,
the result is @xmath347 where @xmath25 is the gauge coupling , which is dimensionful in three dimensions .
when the branes are coincident , the theory is in the symmetric phase and the gauge bosons are massless .
a massless gauge boson gives a contribution of the form @xmath348 which again behaves as a power of @xmath13 .
a dirac fermion of mass @xmath349 yields the contribution @xcite @xmath350 for the massless case , the @xmath13 dependent terms vanish and we have @xmath351 which is a constant independent of the radius @xmath13 . collecting all one - loop contributions , the prefactor in eq.([nucrate ] ) due to the weakly coupled @xmath8 gauge theory in the unbroken phase is given by @xmath352 using ( [ dgoldst ] ) , ( [ dmq]),([da ] ) and ( [ dpsi ] ) we are led to a nucleation rate per unit volume of the form @xmath353 where @xmath354 is the corresponding bounce action for the nucleation of @xmath7 coincident branes .
note that the @xmath13 dependence of the prefactor is simply as a power law .
here we have used the form ( [ dmq ] ) for the scalar contribution @xmath355 , corresponding to an attractive interaction amongst branes which is quadratic at short distances , with a curvature of the potential characterized by some mass parameter @xmath356 . ) and the behaviour changes to @xmath357 , but in any case the dependence is still a power of the radius @xmath13 ] the prefactor in eq .
( [ finald ] ) has the exponential dependence on @xmath263 which counts the number of worldsheet field degrees of freedom , while the euclidean action @xmath354 behaves approximately linearly with @xmath7 .
hence , as suggested in @xcite , the prefactor can be quite important in determining the nucleation rate . in the scenario proposed in @xcite it was also desirable that the enhancement in the nucleation rate would switch off at the present time , in order to prevent the vacuum from decaying further .
unfortunately , the expression ( [ finald ] ) does not seem to have this property .
the prefactor depends only on the radius @xmath13 , which is itself a function of various parameters , such as the brane tension , the charge and the ambient expansion rate , as summarized in table i. according to ( [ finald ] ) , an enhancement of the nucleation rate of coincident branes will occur for @xmath359 where we have used the results of table 1 in the first step .
consider first the situation where @xmath360 .
note that even in the regime when @xmath361 , the dependence of the degeneracy factor on the corresponding ds temperature is only power law , and not exponential as suggested in @xcite . also , it is clear the stability of our vacuum is not guaranteed by the smallness of the present expansion rate .
the enhancement will persist provided that @xmath362 , even if the ambient de sitter temperature vanishes .
more worrisome is the fact that for sufficiently large @xmath7 we enter the regime where @xmath363 . in that case , we have @xmath364 which can get smaller and smaller as we increase the number of coincident branes , eventually leading to a catastrophic decay rate , regardless of the value of @xmath235 .
the expression ( [ finald ] ) for the nucleation rate is valid for @xmath365 . in the opposite limit , @xmath366 , the scalars decouple , contributing a finite renormalization of the parameters in the action ( such as the brane tension and induced newton s constant ) . for completeness ,
this is discussed in appendix a , where it is shown that the renormalization of parameters can have a very significant impact on the nucleation rate . in ref .
@xcite , the prefactor in the nucleation rate for the nucleation of coincident branes was estimated as an entropy enhancement @xmath367 where , from dimensional analysis , the entropy was estimated as @xmath368 per field degree of freedom .
the factor @xmath369 is due to the area of the bubble , and @xmath14 is some effective temperature .
although the interpretation of the prefactor as the exponential of an entropy should not be taken too literally , let us try and phrase the results of the previous subsection in this intuitive language .
a particle detector following a geodesic in a de sitter space responds as if it was at rest in a thermal bath in flat space , at the temperature @xmath370 .
it should be kept in mind , however , that the ds invariant quantum state is in fact a pure state , and hence rather different from a true thermal state .
for instance , any two detectors in geodesic relative motion observe the same temperature , with a perfectly isotropic distribution .
this is a consequence of de sitter invariance , and is in contrast with the situation in a thermal bath in flat space , where moving observers detect a temperature blue - shift in the direction of their motion relative to the bath . the fields living on a nucleated brane will experience some thermal effects too . the bubble is embedded in a ds space characterized by a temperature @xmath22 .
the interior of the bubble is also a ds space characterized by a different expansion rate , with corresponding temperature @xmath16 .
the existence of two different de sitter spaces in contact with the brane led the authors of ref .
@xcite to consider two different possibilities for the effective temperature of the fields on the brane : @xmath371 and @xmath372 . however , there is in fact no ambiguity in the temperature of such fields @xcite , which is determined as follows .
the worldsheet of the brane is an @xmath373 of radius @xmath13 , the euclidean de sitter space in 2 + 1 dimensions . if interactions with bulk fields are neglected , brane fields are only sensitive to the geometry of the worldsheet , and do not know about the properties of the ambient space . in this approximation ,
the relevant temperature is clearly the intrinsic temperature of the lower dimensional de sitter space , @xmath374 this conclusion remains unchanged when we include interactions with bulk fields .
the simplest way to see this , is to consider the limiting case where gravity can be ignored and the nucleation takes place in a flat space .
there , the ambient temperature vanishes @xmath375 , but the fields on the brane will feel the temperature @xmath376 , because the nucleated brane expands with constant acceleration @xmath377 .
an accelerating observer in the minkowski vacuum will detect a rindler temperature @xmath378 , which happens to coincide with the intrinsic worldsheet temperature .
hence , the de sitter vacuum in the 2 + 1 dimensional worldsheet is in equilibrium with the minkowski vacuum in the bulk .
this conclusion is quite general , and applies also to bubbles nucleating in de sitter .
the cdl instanton has an @xmath379 symmetry under euclidean rotations .
this becomes an @xmath380 symmetry after analytic continuation into lorentzian time .
the quantum state after bubble nucleation is expected to inherit this symmetry @xcite , and the only way to achieve it is if the fields on the brane are in their intrinsic de sitter vacuum , which is characterized by temperature @xmath376 .
note that @xmath376 is a relatively high temperature .
the radius of the instanton is always smaller than @xmath231 ( see table 1 ) , and therefore @xmath376 is strictly larger than @xmath22 and @xmath16 .
nevertheless , the product @xmath381 , and hence the entropy enhancement " , is independent of the ambient de sitter temperature .
as shown in the previous subsections , the independence of the prefactor @xmath212 on the ambient expansion rate is only approximate , due to the anomalous behaviour of light fields in de sitter space .
this introduces a dependence of the effective action @xmath382 on the radius @xmath13 of the instanton ( which in turn may depend on @xmath235 in certain regimes ) .
this dependence , however , is quite different from the one proposed in @xcite , where it was suggested that the nucleation rate of coincident branes would be enhanced at high @xmath235 , and would switch off at low @xmath235 due to the drop in ambient temperature .
what we find instead is that , if the cdl instanton for nucleation of coincident branes really exists , then the corresponding degeneracy factor does not necessarily switch off .
we shall return to a discussion of this point in the concluding section . before that , let us turn our attention to a different instanton , which may be relevant to the fmsw scenario .
euclidean de sitter space is compact in all spacetime directions , and ( as we just discussed ) it behaves in some respects as a system at finite temperature .
one may then ask whether there are instantons similar to the thermal ones in flat space .
these correspond to static bubbles , in unstable equilibrium between expansion and collapse .
static instantons with o(3 ) symmetry have previously been considered in a variety of contexts , notably for the description of false vacuum decay in the presence of a black hole ( see e.g. @xcite and references therein ) .
the particular instanton we shall consider corresponds to pair creation of critical bubbles in de sitter , and to our knowledge it does not seem to have received much attention in the past .
this is perhaps not surprising , since its action is higher than the action for the maximally symmetric cdl instanton .
however , if the cdl instanton does not exist for coincident branes , the static one may turn out to be relevant once the degeneracy factors are taken into account .
the energy of a critical bubble is different from zero , and consequently , the metric outside of the bubble is no longer pure de sitter , but schwarzschild - de sitter ( sds ) .
the instanton is a solution of the euclidean equations of motion , with two metrics glued together at the locus of the wall , which is a surface of constant @xmath383 in the static chart of sds ( see fig .
[ estatico ] ) .
the metric outside is given by @xmath384 where @xmath385 , and @xmath386 the metric inside is given by @xmath387 where @xmath388 corresponding to a de sitter solution .
the constant @xmath147 is determined by the condition that on the brane ( i.e. , at @xmath389 ) the two metrics must agree , which leads to @xmath390^{1/2 } $ ] . [ ] [ r]@xmath391[][r]@xmath13[][r]@xmath392 [ ] [ r]@xmath125 [ ] [ l]@xmath78 the parameters @xmath393 and @xmath13 depend on @xmath394 , @xmath395 and @xmath396 .
their values are determined by the junction conditions at the brane @xcite , @xmath108 = -4\pi g\sigma_k \gamma _ { ab } , \label{discont2}\ ] ] where @xmath109 $ ] is the difference in the extrinsic curvature @xmath397 on the two sides and @xmath110 is the world - sheet metric .
( [ discont2 ] ) gives rise to the junction conditions , @xmath398=-4\pi g \sigma_k,\quad [ g']=0 , \label{juco}\ ] ] where we have introduced the new function @xmath399 . using eqs .
( [ efeo ] ) and ( [ efei ] ) , we have @xmath400 hence , using ( [ juco ] ) , @xmath401 , and then @xmath402 and @xmath403 are easily obtained from eqs .
( [ cuadradosprima ] ) : @xmath404 from ( [ efeo ] ) and ( [ efei ] ) we have @xmath405 inserting ( [ ges ] ) in ( [ cuadrados ] ) we finally obtain a quadratic equation for @xmath406 .
the solution is @xmath407 ^{1/2},\ ] ] where we have used @xmath408 .
then the parameters @xmath393 and @xmath13 are given in terms of @xmath409 by @xmath410 this concludes the construction of the instanton solution for given values of the physical parameters @xmath411 and @xmath236 . the equation @xmath412 has three real solutions for @xmath413 .
one of them , say @xmath414 , is negative and the other two are positive .
the two positive roots correspond to the black hole and cosmological horizons .
we call them respectively @xmath415 and @xmath416 .
therefore we can write @xmath417 of course , in our instanton the horizon at @xmath415 is not present , since the exterior metric is matched to an interior metric at some @xmath418 ( see fig .
[ estatico ] ) . for @xmath419
the metric is just a ball of de sitter in the static chart , and it is regular down to the center of symmetry at @xmath420 . in general , the size of the cosmological horizon is given by @xmath421 where we have introduced the angle @xmath422 in the limit @xmath423 the angle @xmath424 , and @xmath425 . according to eq ( [ ges ] ) , at the brane we have @xmath426 , so the equation @xmath427 has a double zero instead of two different roots .
this means that the radius of the instanton will coincide with the radius of one of the horizons only in the special case where both horizons have the same size , @xmath428 . in the limit @xmath429 the exterior
metric becomes the nariai solution @xcite , which has @xmath430 .
note that in the limit @xmath431 , @xmath432 and @xmath433 , as expected . like in the case of instantons describing the production of black holes @xcite or monopoles @xcite in de sitter
, the instanton presented here describes the creation of _ pairs _ of bubbles .
as we shall see , the euclidean solution is periodic in the time direction , so that time runs on a circle @xmath434 ( see fig .
the geometry at the time of nucleation is obtained by slicing the compact instanton through a smooth spacelike surface which cuts the @xmath434 factor at two places , say , @xmath435 and @xmath436 .
the resulting geometry contains two different bubbles separated by a distance comparable to the inverse expansion rate . in order to calculate the temperature of the worldsheet in the static instanton
we must first find the time periodicity @xmath150 .
this is determined by the regularity of the euclidean metric at the cosmological horizon . for @xmath437 , we have @xmath438 where @xmath439 in terms of the new coordinates @xmath440
the metric ( [ metrico ] ) for @xmath441 reads @xmath442 so it is clear that @xmath443 is an angle , @xmath444 , and @xmath125 varies in the range @xmath445 .
therefore the value of @xmath446 is @xmath447 the temperature of the worldsheet instanton is given by the proper time periodicity @xmath448 .
hence , the inverse temperature is given by @xmath449 we shall also be interested in the euclidean action , which turns out to have a rather simple expression in terms of @xmath391 .
this is derived in appendix b , where it is shown that the difference in euclidean actions between the instanton and the background solutions is given by @xmath450 let us start with the case of low tension branes , @xmath451 . in this case the parameter @xmath409 is large compared with @xmath395 , @xmath452 is small , and @xmath453 . in this limit the angle @xmath89 in eq .
( [ varphi ] ) is close to @xmath454 and @xmath455 .
we have @xmath456 this is just the flat space expression for the energy of a critical bubble , multiplied by euclidean time periodicity of the low curvature de sitter space in which this bubble is embedded .
next , we may consider the case of intermediate tension @xmath457 . in this case ,
@xmath458 , @xmath459 , @xmath460 , with @xmath453 , and we have @xmath461 in this case , the difference in pressure between inside and outside of the brane is insignificant compared with the brane tension term , which is balanced against collapse by the cosmological expansion .
finally , in the limit of very large @xmath394 we find that @xmath462 becomes larger than @xmath13 , namely , @xmath463 .
this means that @xmath464 vanishes for some value @xmath465 , given below in eq .
( [ sigmamax ] ) , so it is not sensible to consider the limit of very large @xmath394 but just the limit @xmath466 . as we have mentioned
[ see the discussion below eq .
( [ fofactors ] ) ] , the exterior metric in this limit corresponds to the nariai solution , with @xmath467 . replacing this value in ( [ b ] ) we find readily @xmath468 it is interesting to compare this value of @xmath249 with the corresponding one for the nucleation of black holes in the same de sitter universe .
this is described by the nariai instanton @xcite , which has the bounce action @xmath469 note that the difference @xmath470 , where @xmath471 is the area of the black hole horizon in the nariai solution .
hence , the probability of nucleating black holes divided by the probability of nucleating brane configurations characterized by the same mass parameter is just the exponential of the black hole entropy , as expected from general considerations ( in this argument , we are of course neglecting the entropy stored in the field degrees of freedom living on the branes , which will only show up when the determinantal prefactor in the nucleation rate is evaluated ) .
let us consider the value of @xmath472 in the limit @xmath473 .
this is a singular limit in eq .
( [ betar ] ) due to the simultaneous vanishing of numerator and denominator .
thus we will need to change to more appropriate coordinates .
the fact that @xmath474 does not mean , though , that both horizons coincide , since the coordinates @xmath475 become inadequate in this case . near this limit
the metric outside takes the form ( [ metrico ] ) , with @xmath476 } + r_+^{2}d\omega ^{2}\ ] ] and @xmath477 , plus higher orders in @xmath478 .
the constant @xmath478 is the same parameter defined in ( [ a ] ) , but in this limit tends to zero , @xmath479 .
now we define new coordinates @xmath77 and @xmath480 by @xmath481 so that the metric becomes @xmath482 in these coordinates the cosmological horizon is at @xmath483 and the black hole horizon is at @xmath484 . now in the limit @xmath485 we just replace @xmath486 .
we must determine the position @xmath487 of the brane , which is given as before by the matching conditions ( [ discont2 ] ) , where now the metric outside is ( [ nariai ] ) .
so , on the brane , we have @xmath488 and the extrinsic curvature on the outside of the brane is @xmath489 , with @xmath490 and @xmath491 , i.e. , @xmath492 the curvature inside is as before @xmath493 and @xmath494 , with @xmath495 , so the israel conditions give @xmath496 these equations are easily solved and give @xmath497 so @xmath396 must be less than @xmath498 .
now regularity at the cosmological horizon @xmath499 implies that @xmath500 , so @xmath501 .
hence , @xmath502 thus , also in this case , the effective temperature of the field degrees of freedom living on the worldsheet will be of order @xmath503 or higher .
in flat space , and at finite temperature @xmath504 , the instanton which is relevant for vacuum decay is static and spherically symmetric in the spatial directions .
the fluctuations are periodic in euclidean time , with periodicity @xmath505 .
the worldsheet of the brane has the topology @xmath506 , where the @xmath434 is the direction of imaginary time , and the @xmath507 is the boundary of the
critical " bubble , a closed brane in unstable equilibrium between expansion and collapse ( this is in contrast with the zero temperature instanton , where the world - sheet is a 3-sphere . ) the radius of the critical bubble is given by @xmath508 and the difference of the instanton action and the action for the background is given by @xmath509 where @xmath510 is the classical energy of the critical bubble .
the one loop quantum correction can be written as ( see e.g. @xcite ) @xmath511 here , @xmath512 denotes the free energy and @xmath513 is the correction to the energy of the critical bubble due to the field @xmath260 .
this includes the zero point energy of @xmath260 in the presence of the bubble , as well as the thermal contributions .
finally , @xmath514 is the entropy .
thus , the nucleation rate takes the form @xmath515 where @xmath516 is the total energy , @xmath517 is the total free energy and @xmath518 is the total entropy .
thus , for thermal instantons the determinant prefactor does indeed include the exponential of the entropy .
this is , however , not the only role of the prefactor , since there is also some correction to the energy of the bubble .
consider , for simplicity a massless field @xmath260 . at sufficiently high temperature @xmath519
, the entropy behaves as @xmath520 . from @xmath521
, it follows that @xmath522 .
hence @xmath523 in this case , the prefactor clearly represents an entropy enhancement " .
on the other hand , at lower temperatures , it is not clear whether the prefactor represents an enhancement or a suppression . at temperatures comparable to @xmath524
, the vacuum energy term can be as important as the thermal contributions , and the logarithm of the prefactor @xmath212 can have either sign .
the case of the static instanton in de sitter space is somewhat close to this low temperature situation , and without an explicit calculation it is not clear whether the prefactor represents an enhancement or a suppression of the nucleation rate .
an interesting possibility would be that at sufficiently high gibbons - hawking temperature @xmath525 , the thermal contribution may be sufficient to restore the symmetry , creating the desired attractive force amongst the branes .
this is currently under research .
the static instanton presented in this section may perhaps be better suited to the scenario proposed by feng et al .
@xcite , than the coleman - de luccia instanton . ignoring the degeneracy factors ,
the action of the static instanton is always larger than the action of the cdl instanton . in this sense
, it seems to correspond to a subdominant decay channel .
however , as we have discussed in previous sections , it might well be that the cdl instanton for multiple brane nucleation simply does not exist because of the repulsive force amongst the branes .
this does not exclude the possibility that in the case of the static thermal instantons the symmetry is restored at high ambient ( and worldsheet ) ds temperature @xmath526 .
in this situation , the decay through nucleation of coincident branes would only be possible through the static instanton .
symmetry is not restored in the case of the cdl instanton @xcite . ] at low @xmath235 , the thermal contribution might not be sufficient to restore the symmety and stacks of branes may simply not hold together , destroying the possibility of further decay by coincident brane nucleation .
also , the prefactor and the euclidean action have an exponential dependence on the ambient temperature @xmath527 , and can be much suppressed at the present epoch , contributing to the stability of the present vacuum ( in contrast with the coleman - de luccia case ) .
a potentially worrying aspect of coincident brane nucleation in the cdl case is whether the degeneracy factor may grow without bound as we increase the number of branes @xcite .
as we have seen , this will not happen for the case of the static instanton discussed in this section , since nucleation of coincident branes can not involve arbitrarily large @xmath7 . indeed , there is a maximum value of the combined tension of the branes @xmath528 , given in eq .
( [ sigmamax ] ) , beyond which the instanton simply does not exist . in this limit , the metric outside of the branes approaches the nariai solution . from this observation
, we can easily derive a bound on the maximum entropy which can be stored in the stack of branes .
indeed , the static instanton represents a spherical bubble in unstable equilibrium between undeterred expansion or collapse into a black hole .
the entropy can only increase when the stack of branes collapses , and hence the entropy of the coincident branes can not exceed the entropy of the nariai black hole .
the stack of branes with tension @xmath529 has the same radius as the horizon of the nariai black hole , and so , the entropy of the stack of branes is bounded by one fourth of its own area , in natural units .
a corollary is that the nucleation rate of coincident branes in the limit @xmath529 is bounded above by the nucleation rate of nariai black holes of the same mass .
in this paper we have investigated the possibility of coincident brane nucleation by a four - form field , in connection with string motivated scenarios for the neutralization of the effective cosmological constant . in four dimensions , and after the moduli
are stabilized , the branes repel each other at distances larger than the inverse mass of the moduli . at shorter distances
, their interactions will be model dependent , but in the simplest models the branes do not attract at the classical level .
in this situation , it is unclear whether the coleman - de luccia ( cdl ) instanton for nucleation of coincident branes really contributes to the semiclassical decay rate , since it would have too many zero modes and negative modes . assuming that the cdl instanton exists for the nucleation of coincident branes ( that is , assuming an attractive short range interaction amongst the branes in the stack ) , we have investigated the degeneracy factor accompanying the formula for the nucleation rate , due to the large number of worldsheet degrees of freedom .
we have modeled such degrees of freedom by a weakly coupled sym @xmath8 gauge theory , which is unbroken when the branes are coincident .
we find that the degeneracy factor does not depend very strongly on the ambient de sitter ( ds ) temperatures before or after the nucleation event .
rather , it depends only on the radius of the instanton .
hence the degeneracy factors can be quite important even when the ambient ds temperature is as low as it is today .
this may indicate that nucleation of coincident branes via the cdl instanton is in fact impossible , otherwise the present vacuum would immediately decay . if the cdl instanton for coincident branes does not exist , stacks of branes may still nucleate through a static " instanton which represents pair creation of critical bubbles , in unstable equilibrium between expansion and collapse .
this is the analog of the instanton for thermal activation in flat space . despite the absence of a classical attractive force
, the branes could be held together by thermal corrections to the interbrane potential , which tend to favor the symmetric phase ( where branes are on top of each other ) .
the calculation of this thermal effective potential for the static instanton is currently under research .
one may ask whether a similar symmetry restoration may not happen for the cdl instanton . in this case
the calculation has been done in @xcite , where it is shown that the one loop potential does not help restoring the symmetry .
so it is conceivable that the branes may stick together for the static instanton but not for the cdl instanton , in which case the former would be the relevant decay channel . to conclude
, we have presented some evidence that the saltatory " relaxation scenario of @xcite may be difficult to implement via the cdl instanton , since saltation would be hard to stop at present .
rather , we have speculated that it may be easier to implement through the static instanton . in the scenarios proposed in ref .
@xcite for the saltatory relaxation of the cosmological constant , two different possibilities were suggested for the effective temperature of the worldsheet degrees of freedom , namely @xmath530 and @xmath531 , where @xmath236 and @xmath235 are the expansion rates before and after nucleation .
we have shown that for the static instanton , the relevant temperature is comparable to the ambient de sitter temperature @xmath527 before the tunneling .
hence , the nucleation rate of coincident branes would be unsupressed at large ambient de sitter temperature , but exponentially suppressed at present , which is of course desirable .
clearly , many issues need to be addressed before a scenario based on coincident brane nucleation can be used to successfully explain the smallness of the observed cosmological constant .
a considerable advance would be to understand why the large @xmath0 relaxes to the small @xmath5 instead of plunging directly into deep ads space ( the latter jump involves a larger number of coincident branes and would be rewarded by a larger degeneracy factor ) . in ref .
@xcite an explanation was offered , based on a uniquely weak " form of the anthropic principle . as explained in section ii
, any relaxation mechanism requires the gap @xmath532 in the discretuum of @xmath0 not to be much larger than @xmath5 ( otherwise it becomes a problem to understand why , accidentally , there happens to be an allowed vacumm so close to zero , at @xmath533 ) . in @xcite
it was proposed that @xmath534 with @xmath535 , saturating the above requirement .
then the allowed @xmath0 would take values in the sequence @xmath536 .
if we start from a large @xmath0 , then the enhancement of brane nucleation for large @xmath7 favours a jump to the lowest value in the above list which is still compatible with the existence of observers .
fmsw suggested that the value @xmath537 may already be too small for observers to emerge , making the vacuum with the value @xmath5 the favourite destination .
finally , one should try to embed this scenario in a cosmological context , taking into account the restrictions imposed by homogeneity and isotropy .
if unsuppressed saltation happened after inflation , then we would have seen signals of it in the microwave background .
indeed , bubbles which nucleate after thermalization are still rather small at the time of decoupling , and we would see different domains with different values of @xmath0 separated by fast moving stacks of branes , which would presumably cause large perturbations in the gravitational potential . hence , saltation should occur during inflation , and switch off somewhat before the end of it .
this may impose certain constraints on the space of parameters such as the tension and charges of the branes , or alternatively , on the ambient temperature below which the instanton with coincident branes simply does not exist ( e.g. because thermal symmetry restoration is no longer effective ) .
also , it should be clarified what might be the advantages of a saltatory neutralization " scenario over the randomization " scenarios discussed in section ii .
a possible advantage is that saltatory relaxation operates very quickly , and hence it does not require eternal inflation to take place ( as required in the randomization scenarios ) .
a fuller discussion of these issues is left for further research .
j.g . is grateful roberto emparan and alex vilenkin for very useful discussions .
a.m. was supported by a postdoctoral fellowship of the ministerio de educacin y cultura , spain .
the work by j.g . was supported by cicyt research projects fpa2002 - 3598 , fpa2002 - 00748 , and dursi 2001-sgr-0061 .
the expression ( [ finald ] ) for the nucleation rate is only valid for @xmath365 . for completeness , here we shall discuss the limit @xmath366 . in this case the heavy degrees of freedom decouple , and they simply contribute a finite renormalization of the parameters in front of different operators in the classical lagrangian . for scalar fields we have @xmath538
as discussed in section vii , the first term in this expansion corresponds to a finite renormalization of the brane tension , multiplied by the worldvolume of the stack of branes .
the second term correspond to a finite renormalization of the coefficient in front of the worldsheet ricci scalar .
this term was not present in the classical action we started with , but evidently it can be generated by quantum corrections .
the scalar contribution ( [ effacscalar ] ) tends to decrease the tension of the stack of branes .
this tends to favour the nucleation rate at large @xmath539 , as represented in fig .
[ prefactor ]
. however , whether an actual enhancement really occurs will be model dependent , since all massive species , and not just the scalars , contribute finite renormalizations of the parameters in the action . in section vii
we have assumed that there is an attractive short range force amongst the branes , which we have modeled as a mass term for the scalars representing the relative positions of the branes .
one may expect that the same mechanism which generates a potential for the scalars may generate masses also for their fermionic partners . from ( [ wpsi ] ) ,
heavy fermions give a contribution to the effective potential of the form @xmath540 per degree of freedom .
this gives a positive renormalization of the brane tension ( which tends to suppress the nucleation rate ) .
in the special case where @xmath541 , the brane tension does not renormalize , but each pair of fields will still contribute a finite renormalization of the subleading term @xmath542 which would suppress the nucleation rate .
generically , however , scalars and fermions may wind up with different masses ( since supersymmetry is broken at some level ) , and the tension will be renormalized . dividing the leading term in @xmath543 by the worldvolume
@xmath544=2\pi^2 r^3 $ ] , each scalar and fermionic degree of freedom contributes a brane tension renormalization of the form @xmath545 similarly , there will be a renormalization of the induced newton s constant @xmath546 on the worldsheet , of the order @xmath547 for each pair of heavy field species .
these changes will modify the instanton solution . for @xmath7 coincident branes
, the number of such fields grows as @xmath263 , and the effect of these finite renormalizations can be quite dramatic .
the nucleation rate will take the form @xmath548 where @xmath549 is a constant which depends on whether some ( or all ) of the fermionic species have decoupled or not .
the bounce action @xmath550 in the exponent is calculated by using the renormalized values of the parameters .
if @xmath551 , then the renormalized tension @xmath552 sharply decreases for large @xmath7 , leading to unsuppressed tunneling rate . on the contrary , for
@xmath553 the nucleation of coincident branes is strongly suppressed .
the action is given by @xmath554 on shell , the scalar curvature is given by , @xmath555 and hence the instanton action is given by @xmath556 the first integral in ( [ actionstatic ] ) is just the volume of a two - sphere of radius @xmath13 times @xmath557 .
the second integral in ( [ actionstatic ] ) splits into two contributions from the two regions , @xmath558 so the instanton action is @xmath559 after some algebra @xmath560 can be written in the simple form @xmath561 the exponent @xmath249 which gives the probability for brane nucleation is the difference in euclidean actions between instanton and background .
the action of the background is just @xmath562 so the difference in euclidean actions leads to eq .
( [ b ] ) .
j. j. halliwell , phys .
b * 185 * ( 1987 ) 341 .
s. kachru , r. kallosh , a. linde and s. p. trivedi , phys .
d * 68 * , 046005 ( 2003 ) [ arxiv : hep - th/0301240 ] ; s. kachru , r. kallosh , a. linde , j. maldacena , l. mcallister and s. p. trivedi , arxiv : hep - th/0308055 .
s. b. giddings , phys . rev .
d * 68 * ( 2003 ) 026006 [ arxiv : hep - th/0303031 ] .
j. m. maldacena and c. nunez , int .
j. mod .
phys . a * 16 * ( 2001 ) 822
[ arxiv : hep - th/0007018 ] .
p. candelas and s. weinberg , nucl .
b * 237 * ( 1984 ) 397 .
n. kaloper , j. march - russell , g. d. starkman and m. trodden , phys .
* 85 * ( 2000 ) 928 [ arxiv : hep - ph/0002001 ] .
y. p. goncharov and a. a. bytsenko , class .
* 8 * , l211 ( 1991 ) ; d. muller , h. v. fagundes and r. opher , phys .
d * 66 * ( 2002 ) 083507 [ arxiv : gr - qc/0209103 ] .
r. basu , a. guth and a. vilenkin , phys . rev . *
d44 * , 340 ( 1991 ) .
j. garriga and a. vilenkin , phys .
d * 44 * ( 1991 ) 1007 ; phys .
d * 45 * ( 1992 ) 3469 ; j. guven , phys .
d * 48 * ( 1993 ) 4604 [ arxiv : gr - qc/9304032 ] .
t. vachaspati and a. vilenkin , phys .
d * 43 * ( 1991 ) 3846 .
x. montes , int .
j. theor .
* 38 * ( 1999 ) 3091 [ arxiv : gr - qc/9904022 ] .
v. a. berezin , v. a. kuzmin and i. i. tkachev , phys .
scripta * t36 * ( 1991 ) 269 .
d. a. samuel and w. a. hiscock , phys .
d * 44 * ( 1991 ) 3052 .
p. ginsparg and m. j. perry , nucl .
b * 222 * ( 1983 ) 245 .
h. nariai , science reports of the tohoku univ . 35
( 1951 ) 62 . | nucleation of branes by a four - form field has recently been considered in string motivated scenarios for the neutralization of the cosmological constant .
an interesting question in this context is whether the nucleation of stacks of coincident branes is possible , and if so , at what rate does it proceed .
feng et al .
have suggested that , at high ambient de sitter temperature , the rate may be strongly enhanced , due to large degeneracy factors associated with the number of light species living on the worldsheet .
this might facilitate the quick relaxation from a large effective cosmological constant down to the observed value . here , we analyse this possibility in some detail . in four dimensions , and after the moduli are stabilized , branes interact via repulsive long range forces .
because of that , the coleman - de luccia ( cdl ) instanton for coincident brane nucleation may not exist , unless there is some short range interaction which keeps the branes together .
if the cdl instanton exists , we find that the degeneracy factor depends only mildly on the ambient de sitter temperature , and does not switch off even in the case of tunneling from flat space .
this would result in catastrophic decay of the present vacuum .
if , on the contrary , the cdl instanton does not exist , coindident brane nucleation may still proceed through a static " instanton , representing pair creation of critical bubbles a process somewhat analogous to thermal activation in flat space . in that case , the branes may stick together due to thermal symmetry restoration , and the pair creation rate depends exponentially on the ambient de sitter temperature , switching off sharply as the temperature approaches zero
. such static instanton may be well suited for the saltatory " relaxation scenario proposed by feng et al . |
during a ccd imaging campaign with the 40-inch telescope at siding spring observatory carried out by one of us ( hj ) in 1999 , the galaxy ngc 5011c only 7.7 arcmin away from the target galaxy eso 269-g070 was falling into the @xmath6sqarcmin field - of - view and attracted the attention because of its morphological appearance resembling closely that of a nearby dwarf elliptical galaxy .
most strikingly the stellar system exhibits the featureless light distribution with the characteristic low central surface brightness .
however , a search through the literature for more details about this galaxy suggested a different picture .
based on its position in the sky and published redshift measurements of over 3000kms@xmath7 , ngc 5011c should be a member of the distant centaurus cluster .
if true , the immediate question arises what kind of galaxy this is with a physical size larger than that of the neighboring centaurus cluster member , the s0 galaxy ngc 5011b , but with the stellar surface density of a low luminous dwarf galaxy ( see fig .
1 ) ? the galaxy ngc 5011c ( eso
269-g068 ) is located in the general direction of the cena group and the centaurus cluster .
the nearby cena group has a mean distance from the milky way of 3.7 mpc ( jerjen , freeman & binggeli @xcite ; karachentsev et al .
@xcite ) and a mean heliocentric velocity @xmath8kms@xmath7 ( ct et al .
@xcite ) . on the other hand ,
the centaurus cluster is @xmath9 times further away at 45.3mpc ( mieske , hilker & infante @xcite ) .
dickens , currie & lucey ( @xcite ) reported a redshift of 3195kms@xmath7 for ngc 5011c ( galaxy # 531 ) , consistent with the 3200kms@xmath7 from fairall ( @xcite ) , which immediately implies an association with the cen30 component of the centaurus cluster that has a mean velocity of @xmath10kms@xmath7 ( lucey , currie & dickens @xcite ) . from its location ,
ngc 5011c seems to form a physical pair with the other centaurus cluster member , the s0 galaxy ngc 5011b ( eso 269-g067 ) , which is only 1.1arcmin to the north ( fig . 1 ) and has a very similar redshift of 3294kms@xmath7 as reported by the same authors ( galaxy # 530 ) .
however , despite the small projected distance of only 14kpc there is no obvious optical sign of interaction between the two . more confusion was added to the picture by the results from dressler ( @xcite ) .
the galaxy catalog listed a redshift only for ngc 5011c ( 3293kms@xmath7 , galaxy # 456 ) while no measurement was reported for ngc5011b .
this appeared somewhat strange as the central surface brightness of ngc5011b is approximately four magnitudes higher than that of ngc5011c making the former galaxy a more logical target for a redshift measurement .
the present paper reports on our spectroscopic and optical follow up of the galaxy pair to shed more light on this confusing situation around the two galaxies ngc5011b and c. in the next section we describe our 3.6 m observations and data reduction , and the age and metallicity of the stellar populations dominating the galaxies light are determined in 3 .
the new recession velocities for the two galaxies are presented in 4 followed by a photometric analysis with a rough distance estimate for ngc5011c in 5 and a discussion in 6 .
the observations were carried out with the eso 3.6 m telescope at la silla , during an efosc2 ( buzzoni et al .
@xcite ) technical night , with the purpose of testing all science and calibration templates , and to estimate the photometric zero points and the grism efficiencies .
the `` rotate to slit '' template was tested by putting the nuclei of ngc 5011b and c into the same 1arcsec slit , and taking @xmath11sec spectra .
before and after the galaxies observation , spectrophotometric standards were observed , and immediately after the end of the last spectrum , @xmath12sec spectra of a radial velocity standard were also taken .
although several grisms were tested during the night , for this particular set of data grism # 11 was used , since it gives a good wavelength coverage and its intermediate dispersion allows to obtain a good s / n for faint targets in relatively short times .
an observing log is listed in table [ tabcap : journal - of - the ] .
the radial velocity standard is hd125184 ( from udry et al .
@xcite ) , and the rest of the objects in the table are spectrophotometric standards from hamuy et al .
( @xcite ) .
all reductions were carried out with the eso - midas data processing system ( e.g. warmels @xcite ) , and in particular within the long context , which provides commands to perform the standard reduction steps of long - slit spectroscopy .
a series of scripts were written to facilitate as much as possible the batch reduction of several spectra , and they are publicly available in the efosc2 instrument pages .
we first prepared the system to reduce the night spectra , using the afternoon calibrations : the frames were rotated to have the wavelength increasing along the ccd rows , a median bias frame was created and subtracted from the rest of the frames , and a normalized flat - field was also generated . to do this , five flat - fields taken with slit @xmath13 were median combined , and then a two - dimensional polynomial was fitted to the resulting slit .
a database of wavelength solutions for each efosc2 grism is also published in the instrument web pages , so by using the appropriate guess session for grism # 11 it was straightforward to get the solution for the current data set . after searching the master arc for emission lines , and after defining a two - dimensional grid of points distributed on the arc ridge lines ,
the grid was fitted with a two - dimensional polynomial , with degree 3 along the dispersion axis , and degree 2 along the spatial direction , which is enough to correct the curvature of the arc lines .
the r.m.s .
error of the wavelength solution is @xmath14 , and the calibrated spectra cover the range 3383 to 7505 with a dispersion of 4.23 / pix .
the width of the arc lines at 5000 is @xmath15 , so a velocity measurement based on a single spectral line has a 1-sigma error of @xmath16 . during the night , after a new spectrum was taken , it was reduced using the calibration data prepared above : it was rotated and trimmed , bias - subtracted and flat - fielded , and the wavelength calibration was applied , which transforms the bidimensional frames from the pixel @xmath17 space to the @xmath18 space , where @xmath19 is the spatial direction after correction for the distortion introduced by the grism .
the one - dimensional spectra were extracted with a simple average over the rows covered by the profile , and then multiplied by the number of rows .
the extraction window was defined by inspecting the spatial profile of the spectra , and taking the rows were the spectrum was significantly brighter than the background .
prior to the extraction , the sky background was subtracted from each frame : two template sky spectra were defined in two windows preceding and following the target spectrum , and the sky background at the position of the target was defined by a linear interpolation between these two spectra .
the spectrophotometric standard ltt 3218 was used to define the response function of the grism : in this case the spectrum was extracted using a very wide window and then compared to the spectrum published by hamuy et al .
( @xcite ) .
the spectra were flux - calibrated by first normalizing by their exposure time and then dividing by the response function . prior to this
, they were corrected for atmospheric extinction using the average curve at la silla .
the fully reduced and calibrated spectra of the two galaxies are shown in fig .
[ figcap : the - flux - calibrated - spectra ] .
the spectral energy distribution ( sed ) of ngc 5011b is obviously redder than that of ngc 5011c , but a more quantitative comparison of these two spectra is not possible , since the slit was aligned along the two galaxy centers , and not along the parallactic angle . to overcome this problem , we started from the fact that ngc 5011b is classified as an s0 galaxy ( lauberts @xcite ) . looking at fig . 1 of zaritsky et al .
( @xcite ) one can see that the seds of this class of galaxies are rather uniform , so by selecting a template spectrum and comparing it to our observed one , we can estimate the degree of loss of blue flux induced by the differential atmospheric refraction .
zaritsky et al .
used the spectral atlas of kennicutt ( @xcite ) , which includes five s0 galaxies .
we chose ngc 3245 since , compared to the other templates , it has no emission lines , its sed is not rising in the blue , it extends to redder wavelengths , and it has better spectral resolution . by dividing the ngc 5011b spectrum by the template (
redshift corrected ) , a correction function was found , which is smoothly declining from red to blue , reaching at @xmath20 @xmath21 of the value at @xmath22 . in other words ,
@xmath23 of the blue flux was lost during the observation .
the trend of the correction is well fit by a straight line , which was then used to compute the intrinsic seds of the two galaxies .
the corrected spectra are shown in fig .
[ figcap : the - flux - calibrated - spectra ] together with the original ones , and after normalizing their fluxes to unity at @xmath24 .
the original spectra of ngc 5011b and ngc 5011c have fluxes reaching a maximum of @xmath25 erg @xmath26 sec@xmath7 @xmath7 in the @xmath27 band and @xmath28 erg @xmath26 sec@xmath7 @xmath7 in the @xmath29 band , respectively . in the @xmath27 band ,
ngc 5011c is four magnitudes fainter than ngc 5011b .
a first check of the correctness of our seds can be done by using them to perform synthetic broad - band photometry in order to compute colors , and then comparing the result to that obtained with our direct imaging . to compute synthetic colors ,
a stellar a0i template was taken from jacoby et al .
( @xcite ) library , and instrumental magnitudes and color were computed by integrating it within the johnson @xmath29 and @xmath30 filter passbands from bessell ( @xcite ) .
a zero - point correction was then computed , to calibrate the color to the standard johnson system ; the color of an a0i star ( @xmath31 ) was taken from allen s astrophysical quantities ( cox @xcite ) . by repeating this procedure with the two galactic seds as input , we found @xmath32 and @xmath33 for ngc 5011b and ngc 5011c , respectively .
the spectrum of ngc 5011c has been extracted in the central @xmath34 , and in fig .
[ fig : sbprofiles ] one can see that @xmath35 at the center of the galaxy ( see below ) .
the synthetic color is thus in good agreement with that obtained from broad - band surface photometry , which means that our correction indeed yields a reliable sed .
after this consistency check , we can interpret the two seds using population synthesis techniques .
for that purpose we used galaxev from bruzual & charlot ( @xcite ; bc03 ) , to estimate the luminosity - weighted age and metallicity
. a grid of model seds were generated , based on the padova 1994 tracks ( see bc03 for the references ) with an input salpeter stellar mass function .
computations were done for ages ranging from @xmath36 gyr to @xmath37 gyr , and for five of the available metallicities @xmath38 , @xmath39 , @xmath40 , @xmath41 , and @xmath42 . the age was increased in steps of @xmath36 gyr up to @xmath43 gyr , and in steps of @xmath43 gyr for older ages , and the spectra were degraded to our instrumental resolution with software provided by bc03 .
the best - fit was searched by minimizing the difference between model and observed seds . in this way we found that the best solution is obtained for an age of @xmath4 gyr and @xmath44 for ngc 5011c , and for an age of @xmath3 gyr and @xmath45 for ngc 5011b .
in the bc03 scale the metallicity of the sun is @xmath45 , so the two metallicities are @xmath5 solar and solar , respectively .
the good match between the theoretical and observed seds can be checked in fig .
[ figcap : the - flux - calibrated - spectra ] .
note that these results only apply to the central regions of the two galaxies ; for example the ngc 5011c spectrum was extracted in the central @xmath34 , which correspond to @xmath46 pc at the estimated distance of the galaxy ( see below ) .
the age and metallicity thus represent the central stellar population only , and since the rest of the galaxy has redder colors presumably it hosts older populations , as is evident from fig .
[ fig : sbprofiles ] .
indeed a kind of tiny , whiter , bulge - like core can be seen in fig .
[ figcap : efosc2-true - color - image ] as well . the total @xmath30 magnitude computed in [ sec : surface - photometry ]
allows us to estimate the mass of ngc 5011c as well . at a fiducial distance of @xmath47
mpc it converts into @xmath48 , and assuming @xmath49 , the total luminosity can be computed . finally , from fig .
1 of bc03 one can estimate that , for an age of @xmath43 gyr , the mass - to - light ratio is @xmath50 , so the mass of the galaxy is @xmath51 .
this value of the mass is only valid in the single age approximation , but it s very likely that the sfh of the galaxy is more extended , so realistically that mass represents only a lower limit .
but even with this caveat , the stellar mass is in the regime of that of dwarf galaxies .
a brief inspection of the prominent balmer lines in the spectra ( see fig . [ figcap : the - flux - calibrated - spectra ] ) quickly reveals that the two galaxies must have quite different redshifts in contrast to what has been reported in the literature .
an initial estimate of the recession velocities can be obtained by measuring the central wavelength of one of these lines .
for example , the center of h@xmath52 is at @xmath53 in the ngc 5011b spectrum but at @xmath54 in the ngc 5011c spectrum ( see also fig . [
fig : the - upper - two ] ) .
this translates into a velocity of @xmath55 for ngc 5011b while that of ngc 5011c would be @xmath56 , where the error is the minimum uncertainty estimated above at @xmath57 . to compute more precise velocities ,
we cross - correlated our spectra with that of the radial velocity standard hd125184 , thus obtaining the relative velocity of each galaxy with respect to that of the standard star . if we call this relative velocity @xmath58 , then the true heliocentric velocity of the galaxy is@xmath59 where @xmath60 is the heliocentric correction listed in table [ tabcap : journal - of - the ] , and the heliocentric velocity of the standard star is @xmath61 . from the same table
, one can see that @xmath62 , where the individual heliocentric corrections have been computed using iraf s task rvcor .
the cross - correlation of the spectra has been done using iraf s task fxcor . to improve the signal ,
the spectra were first trimmed to @xmath63and then continuum - normalized , as shown in fig .
[ fig : the - upper - two ]
. additionally , in the case of ngc 5011c the region containing h@xmath64 and the g band was masked out before cross - correlating the two spectra : the two features are located close to each other at our spectral resolution and as h@xmath64 is basically absent in the rv standard spectrum , the algorithm would overestimate the velocity as it attempts to correlate h@xmath64 in the galaxy spectrum with the g band in the standard star spectrum . masking
the two spectra accordingly removed the confusion leading to a more realistic velocity estimate with a reduced measurement error . at the end of this analysis
we get @xmath65 and @xmath66 , so the heliocentric velocities are @xmath2 and @xmath1 for ngc 5011b and ngc 5011c , respectively .
to gain insight into ngc5011c s colors and photometric parameters , shallow broad - band b , v , and r frames of 90s , 60s , and 30s integration time , respectively , were acquired with the efosc2 instrument during the same observing night ( see table [ tabcap : journal - of - the ] ) .
a montage image is shown in fig .
[ figcap : efosc2-true - color - image ] .
the two galaxies have similar angular size , but they are of entirely different morphological type .
ngc 5011b is a classical lenticular ( s0 ) galaxy seen almost edge - on , while ngc 5011c appears as a low surface brightness , early - type system with a uniform color and a hint of a tiny nucleus or central star concentration .
continuous meteomonitor measurements of the atmospheric conditions at la silla observatory showed that the extinction rms was less than 0.02 magnitudes indicating that the sky transparency was good . from 911 measurements spread in time the photometric zero points for the three passbands
were calculated : zp(b)@xmath67 , zp(v)@xmath68 , and zp(r)@xmath69 .
the pre - reduction of the raw ccd frames was carried out within midas in the usual manner .
after the bias subtraction , pixels that do not respond linearly to flux were replaced by a median of the surrounding area , using the latest efosc2 bad pixel mask , and then each science frame was flattened using high signal - to - noise sky flats .
all reduced frames were flat to less than 0.08% .
the sky level was determined from star - free regions well away from the galaxies .
then foreground stars and background galaxies were removed from the images to allow uncontaminated photometric measurements using procedures written within the iraf package .
stars in the vicinity of ngc5011c were replaced by small patches of empty sky .
the area underneath the neighbor galaxy ngc5011b was restored using a rectangular sub - image from the opposite side of the galaxy .
the center of ngc5011c was given by the intensity - weighted centroid .
the galaxy flux was integrated using circular apertures on the clean @xmath70 images to produce growth curves as a function of the geometric radius @xmath71 ( where @xmath72 and @xmath73 are the major and minor axis ) .
the total magnitude @xmath74 was determined by extrapolation to the asymptotic intensity , and the effective radius @xmath75 was defined as the value at which the intensity reached half of @xmath74 . within @xmath75
the mean surface brightness @xmath76 was also computed .
these photometric parameters are listed in table [ photdata ] .
surface brightness profiles in the three bands and extinction - corrected colour gradients ( b r)@xmath77 , ( v r)@xmath77 , and ( b v)@xmath77 are shown in figure [ fig : sbprofiles ] .
thereby we used the extinction values a@xmath78 , a@xmath79 , and a@xmath80 from schlegel , finkbeiner & davis ( @xcite ) .
additionally , srsic parameters were measured by fitting a srsic model ( srsic @xcite ) : @xmath81 to the observed profiles .
these model parameters are also listed in table [ photdata ] and the corresponding model curves superimposed onto the surface brightness profiles in fig . [ fig : sbprofiles ] .
the quoted uncertainties include formal errors owing to the sky background removal , photometric uncertainty and the fit to the surface brightness profile .
the total magnitudes are in good agreement with the values published in the eso - uppsala catalogue ( lauberts & valentijn @xcite ) . although the current imaging data is not deep enough to measure an accurate distance for ngc5011c by means of the surface brightness fluctuations method ( e.g. jerjen et al .
@xcite ) we can make a guess at the distance interval in which ngc5011c is likely to be found with the help of the two independent quantities @xmath76 and @xmath82 if we accept the galaxy to be of early - type morphology .
\(a ) the @xmath83 relation for early - type dwarf galaxies in the virgo cluster ( binggeli & cameron @xcite ) predicts an apparent b magnitude of @xmath84mag for a galaxy like ngc5011c with an extinction - corrected mean effective b surface brightness of @xmath85 mag arcsec@xmath86 at the virgo distance . employing the latest cluster distance modulus of @xmath87mag ( jerjen , binggeli & barazza @xcite , mei et al .
@xcite ) this translates into an absolute magnitude of @xmath88mag . combining that number with the apparent b magnitude in table [ photdata ] we infer a distance modulus of @xmath89mag or a distance range of 4.718.6mpc .
\(b ) as an alternative , we can use the srsic parameter @xmath82 that controls the overall shape of a galaxy to estimate the distance interval .
equation ( 6 ) from binggeli & jerjen ( @xcite ) that describes the @xmath90 relation for early - type dwarfs in the virgo cluster predicts an apparent b magnitude of @xmath91 mag for ngc5011c at the virgo distance . using the same cluster distance modulus as before yields an absolute magnitude of @xmath92 mag , a galaxy distance modulus of @xmath93mag , and a distance range of 4.39.9mpc .
the main conclusion we can draw from these estimates based on photometric properties is that the derived distance intervals are consistent with ngc5011c being a member of the cena group at 3.7mpc ( jerjen , freeman & binggeli 2000 ; karachentsev @xcite ) but incompatible with the centaurus cluster distance of 45.3mpc ( mieske , hilker & infante @xcite ) .
the deep spectroscopy and shallow imaging with the 3.6 m eso telescope has confirmed a new member of the nearby cena group whose true identity remained hidden because of coordinate confusion and wrong redshifts in the literature since 23 years .
the galaxy s morphology provided the deciding clue .
ngc 5011c is only 2.25 degrees to the west of the dominant group galaxy ngc 5128 ( centaurus a ) . adopting a mean group distance of 3.7mpc
this angular separation translates into 150kpc .
the two galaxies have also comparable velocities with @xmath1 for ngc 5011c and @xmath94 for ngc 5128 . the sky distribution of all known group members in the vicinity of ngc 5128 is shown in fig .
[ figcap : distribution ] .
heliocentric velocities and distances taken from the literature ( jerjen , binggeli & freeman @xcite ; karachentsev @xcite ; karachentsev et al .
@xcite ) are given in parentheses .
ngc 5011c is well within the velocity distribution of the cen a group , and it is definitely not compatible with the velocity distribution of the centaurus cluster .
the cen 30 component of the cluster has a velocity dispersion of @xmath95 km sec@xmath7 , so ngc 5011c s velocity is @xmath96 lower than the cluster s average .
although very high peculiar velocities can be observed in galaxy clusters , this is not the case for the centaurus cluster , and all its members have velocities greater than @xmath97 km sec@xmath7 ( see fig . 3 in lucey et al .
@xcite ) . at 3.7mpc , ngc 5011c has an absolute blue magnitude of @xmath98mag that makes it the 8th brightest member of the group comparable in luminosity to eso 269 - 066 and eso 384 - 016 . in terms of photometric properties such as the overall ( b r)@xmath77 color and the srsic parameters the galaxy resembles closely the intermediate type ( ds0/im ) dwarf eso 384 - 016 ( see table 2 of jerjen , binggeli & freeman @xcite ) .
ngc 5011c has a bluer global appearance ( @xmath99b r@xmath100mag ) than ngc 5011b but unlike for instance the blue lsb galaxy ugc 5889 ( vallenari et al .
@xcite ) it shows no sign of on - going star formation neither in the 2d image ( e.g. young star clusters , regions ) nor in our 1d - spectrum ( emission lines ) .
our estimated age of @xmath101 gyr for the central stellar population , and the presumably older disk , is compatible with this scenario , since much younger stars would be needed to ionize any existing gas .
ngc 5011c is neither a compact blue dwarf nor does it exhibit an irregular stellar distribution or spiral arms .
optically , the galaxy has smooth , symmetric isophotes suggesting a classification as dwarf elliptical or intermediate ( de / im ) type galaxy .
a look at the 21 cm spectrum from the hi parkes all - sky survey ( barnes et al .
@xcite ) observed in the direction of ngc 5011c gives the impression of some emission at the optical velocity ( see fig .
[ figcap : hipass ] ) . however , the periodic baseline structure in the velocity range 150kms@xmath102kms@xmath7 makes it impossible to draw a final conclusion on the possible presence or absence of in ngc 5011c . at the tentative distance of 3.7mpc a non - detection corresponds to an upper limit of @xmath103 which is still a factor of 2.5 higher than the hi mass of @xmath104 detected in eso 384 - 016 ( beaulieu et al .
@xcite ) . for a most recent list of detections and upper limits in cen a group
dwarf galaxies see bouchard et al .
( @xcite ) . to obtain a better hi mass estimate and to measure an accurate distance from surface brightness fluctuations or the tip magnitude of the red giant branch
will be part of a follow - up study of ngc 5011c . | a critical study of the properties of groups of galaxies can be done only once a complete census of group members is available . despite extensive surveys
, even nearby groups can lead to surprises .
indeed we report the discovery of a previously unnoticed member of the centaurus a group , ngc 5011c . while the galaxy is a well known stellar system listed with a ngc number its true identity remained hidden because of coordinate confusion and wrong redshifts in the literature .
ngc 5011c attracted our attention since , at a putative distance of @xmath0 mpc , it would be a peculiar object having a very low surface brightness typical of a dwarf galaxy , and at the same time having the size of an early - type spiral or s0 galaxy . to confirm or reject this peculiarity ,
our immediate objective was to have the first reliable measurement of its recession velocity .
the observations were carried out with efosc2 at the 3.6 m eso telescope , and the spectra were obtained with the instrument in long - slit mode .
the redshifts of both ngc 5011c and its neighbor ngc 5011b were computed by cross - correlating their spectra with that of a radial velocity standard star .
we found that ngc 5011c has indeed a low redshift of @xmath1 and thus is a nearby dwarf galaxy rather than a member of the distant centaurus cluster as believed for the past 23 years .
rough distance estimates based on photometric parameters also favor this scenario . as a byproduct of our study we update the redshift for ngc 5011b at @xmath2 . applying population synthesis techniques
, we find that ngc 5011b has a luminosity - weighted age of @xmath3 gyr and a solar metallicity , and that the luminosity - weighted age and metallicity of ngc 5011c are @xmath4 gyr and @xmath5 solar .
finally we estimate a stellar mass of ngc 5011c comparable to that of dwarf spheroidal galaxies in the local group . |
in @xcite it was proved that harmonic polynomials @xmath0 , @xmath5 , have at most @xmath3 complex zeros .
an example was provided that this bound is sharp for @xmath6 and it was conjectured that it is sharp for every @xmath4 .
the main result of this note is the following theorem , which was conjectured by crofoot and sarason ( * ? ? ?
* conj . 1 ) .
[ mainthm ] for every @xmath7 there exists a complex analytic polynomial @xmath1 of degree @xmath4 and mutually distinct points @xmath8 with @xmath9 and @xmath10 .
in particular , this immediately implies the sharpness of the bound @xmath3 for every @xmath2 .
[ maincor ] for every @xmath7 there exists a complex analytic polynomial @xmath1 of degree @xmath4 such that @xmath11 has @xmath3 zeros .
it was proved in @xcite that theorem [ mainthm ] implies corollary [ maincor ] .
however , for the convenience of the reader we will provide an alternative argument to derive this corollary after the proof of the theorem .
the main ingredient in the proof of theorem [ mainthm ] is an explicit construction of a `` topological polynomial '' and a result of levy that certain topological polynomials are thurston - equivalent to polynomials .
in addition we are able to calculate the number of real polynomials @xmath1 satisfying the conditions in theorem [ mainthm ] , up to equivalence .
two polynomials @xmath1 and @xmath12 are _ conjugate _ if there is an affine linear map @xmath13 with @xmath14 .
they are _ equivalent _ if there are affine linear maps @xmath15 and @xmath13 with @xmath16 .
[ countingthm ] the number @xmath17 of equivalence classes of real polynomials @xmath1 of degree @xmath4 having @xmath18 distinct critical points @xmath19 and satisfying @xmath20 for @xmath21 is @xmath22 , where @xmath23 is the @xmath24-th catalan number .
the number of conjugacy classes is @xmath25 if @xmath26 , and @xmath27 if @xmath28 .
there are non - real polynomials satisfying the same equations .
in particular , for any real solution @xmath1 , and @xmath29 on the unit circle , the function @xmath30 is also a solution .
we conjecture that all non - real solutions are equivalent to real solutions .
the proof of theorem [ countingthm ] relies on poirier s results about hubbard trees for post - critically finite polynomials from @xcite and @xcite .
if one allows rational functions @xmath1 instead of polynomials , the number of zeros of @xmath11 is at most @xmath31 ( see @xcite , which also outlines the connection of this problem to gravitational lensing ) , and the proof that this bound is sharp is much easier and very explicit @xcite
. acknowledgment .
i would like to thank dmitry khavinson for helpful discussions .
we will construct a polynomial with real coefficients .
there is a minor difference between the cases where @xmath4 is even and the one where @xmath4 is odd . in the case of even @xmath4 , one of the points will be real , in the odd case all @xmath32 will be non - real .
note also that the polynomial @xmath1 itself will map all @xmath32 to their complex conjugates , so the second iterate fixes all @xmath32 .
these polynomials are special cases of post - critically finite rational maps .
the main ingredient in the proof is thurston s topological characterization of rational maps @xcite , and a result of levy on topological polynomials . for background
see also @xcite and @xcite .
a _ topological polynomial _ is an orientation - preserving branched covering @xmath33 of the sphere to itself with @xmath34 .
critical points are points where @xmath33 is not locally injective , the _ post - critical set _
@xmath35 is the closure of @xmath36 .
an orientation - preserving branched covering of the sphere is _ post - critically finite _ if the post - critical set contains finitely many points , or equivalently if all critical points are eventually periodic under iteration .
two post - critically finite maps @xmath33 and @xmath37 are _ thurston equivalent _ if there exist orientation - preserving homeomorphisms @xmath38 and @xmath39 , homotopic rel @xmath35 , with @xmath40 .
thurston gave a criterion for post - critically finite branched coverings to be equivalent to rational maps , and levy proved the following about topological polynomials ( see @xcite , ( * ? ? ?
* cor . 5.13 ) ) . if @xmath33 is a topological polynomial such that every critical point eventually lands in a periodic cycle containing a critical point , then @xmath33 is thurston - equivalent to a polynomial .
the polynomial is unique up to affine conjugation .
we will first construct a topological polynomial @xmath33 satisfying @xmath41 for all @xmath42 , and @xmath43 for all critical points @xmath44 .
the cases of even and odd degrees are slightly different , so we will start with the case of odd degree @xmath45 for some @xmath46 .
let @xmath37 be a polynomial of degree @xmath47 with real coefficients , @xmath24 distinct critical points @xmath48 in the upper halfplane such that the corresponding critical values @xmath49 are mutually distinct points in the lower halfplane .
e.g. , one could choose a small constant @xmath50 , define @xmath51 , @xmath52 , and @xmath53 it is easy to check that @xmath37 satisfies all the conditions if @xmath54 is sufficiently small . now choose a homeomorphism @xmath55 of the closed lower halfplane with @xmath56 .
extend @xmath55 by reflection to a homeomorphism of the plane and define @xmath57 .
then @xmath33 is a topological polynomial of degree @xmath47 , commuting with reflection in the real line , mapping all its critical points to their complex conjugates .
the post - critical set of @xmath33 consists of all the critical points @xmath58 and @xmath59 . by levy
s result @xmath33 is thurston - equivalent to a polynomial @xmath1 .
the uniqueness part of levy s theorem implies that @xmath1 is conjugate to @xmath60 , so @xmath1 is conjugate to a real polynomial , and we may as well assume that @xmath1 itself is real .
it also follows that both @xmath38 and @xmath39 are symmetric with respect to the real line . by the definition of thurston - equivalence we know that @xmath61 , so @xmath1 has critical points @xmath62 , and @xmath63 for all @xmath64 .
if the degree @xmath4 is even , we have to modify the initial step of the construction .
write @xmath65 with some @xmath66 , and find a real polynomial @xmath37 of degree @xmath4 with distinct simple critical points @xmath67 , @xmath68 in the upper halfplane , such that @xmath49 are all distinct and in the lower halfplane for @xmath69 . here
one can take @xmath70 with @xmath71 chosen as above for @xmath72 . now choose a homeomorphism @xmath55 of the closed lower halfplane onto itself such that @xmath73 and @xmath74 for @xmath72 . the rest of the proof is identical to the case of odd degree .
proof of corollary [ maincor ] .
we will show that @xmath0 has at least @xmath3 zeros . together with the result from @xcite
, this shows that it has exactly @xmath3 zeros .
first we observe that @xmath1 has no degenerate fixed points , i.e. all fixed points of @xmath75 satisfy @xmath76 .
( for holomorphic maps , a fixed point is degenerate if the derivative is exactly 1 , for anti - holomorphic maps it is degenerate whenever the derivative has modulus 1 . )
if this were the case , then @xmath77 would have a parabolic fixed point , which would have a critical point in its attracting basin .
however , all critical points of this maps are mapped to the critical points of @xmath1 and those are fixed .
having only non - degenerate fixed points , the sum of the lefschetz indices of the fixed points in the riemann sphere has to be equal to @xmath78 , as the induced maps on 0- and 2-dimensional homology are the identity and multiplication by @xmath79 , respectively .
it already has super - attracting fixed points at @xmath59 and @xmath80 .
the lefschetz index of a super - attracting fixed point is @xmath81 , thus the sum of the indices of all those fixed points is @xmath4 .
we conclude that @xmath1 must have at least @xmath82 fixed points with index @xmath83 , i.e. repelling fixed points for @xmath60 .
this proves sharpness .
we will use results about hubbard trees by poirier ( see @xcite and @xcite ) in order prove theorem [ countingthm ] .
hubbard trees were first introduced in @xcite in order to classify post - critically finite polynomials without periodic critical points . in that case
the julia set is a dendrite , and the hubbard tree is the embedded subtree of the julia set spanned by the post - critical set .
it has a finite number of vertices and edges .
the polynomial induces a map on the tree , and conversely an embedded finite tree with a map satisfying certain conditions gives rise to a post - critically finite polynomial , unique up to affine conjugation .
poirier generalized this to include the case of post - critically finite polynomials with critical periodic points . in this case
all fatou components are simply connected with a distinguished base point , the unique preperiodic point in the component .
the hubbard tree in this case intersects both the julia and the fatou set , and the part in a fatou set consists of preimages of radial lines under the riemann map mapping the component to the unit disk and the basepoint to 0 .
this construction determines an angle between any two edges meeting at a point in the fatou set . the important information needed to reconstruct the map from the ( marked ) tree is the local degree at every vertex , the angles between any two edges meeting at a fatou vertex , and the mapping on vertices .
there is no well - defined angle between edges meeting at points in the julia set , but there is a well - defined cyclic order , and we define the `` angles '' between the edges to be all the same . in that way we end up with an angle function on the tree defined for any pair of incident edges .
an abstract hubbard tree is a finite tree embedded in the plane , together with a mapping @xmath84 on the vertices , a local degree function @xmath85 on the vertices , and an angle function @xmath86 defined for any pair of incident edges @xmath87 .
it has to satisfy @xmath88 , @xmath89 , and @xmath90 whenever @xmath91 meet at a vertex .
the map @xmath92 satisfies @xmath93 whenever @xmath94 and @xmath95 are adjacent vertices . in this way
@xmath92 induces a mapping on the edges , mapping the edge @xmath96 $ ] homeomorphically to the unique path between @xmath97 and @xmath98 , i.e. mapping an edge to a union of edges . if two edges @xmath99 and @xmath100 meet at @xmath94 , then @xmath101 and @xmath102 meet at @xmath97 , and we require that @xmath103 , which is just the fact that holomorphic maps of local degree @xmath104 multiply angles by @xmath104 .
a vertex is a fatou vertex if it eventually lands in a periodic vertex @xmath94 with @xmath105 .
all other vertices are julia vertices .
the last and not quite as obvious requirement for the hubbard tree is that @xmath92 is expanding , i.e. whenever @xmath94 and @xmath95 are adjacent julia vertices , there exists @xmath4 such that @xmath106 and @xmath107 are not adjacent .
the _ degree _ of the hubbard tree is @xmath108 .
poirier s main result in @xcite is that any such abstract hubbard tree corresponds to a polynomial of degree @xmath109 , unique up to affine conjugation .
the uniqueness part easily implies that any symmetries of the hubbard tree correspond to symmetries of the polynomial .
proof of theorem [ countingthm ] .
we will first prove the result about the number of conjugacy classes . by poirier s
theorem we only have to count the number of corresponding hubbard trees up to symmetry .
we will first establish some general properties of the corresponding hubbard trees , after which we will treat odd and even degrees separately .
the following claim gives a complete characterization of the hubbard trees in question , the rest of the argument is counting . the induced map on the hubbard tree is @xmath110 .
the only vertices in the hubbard tree are the critical points .
all angles are @xmath111 or @xmath112 .
conversely , any finite tree @xmath13 which is symmetric with respect to complex conjugation , has local degree @xmath113 at every vertex , with the map @xmath114 , @xmath115 , arises as a hubbard tree of a polynomial satisfying the requirements in theorem [ countingthm ] .
let @xmath116 and @xmath117 be critical points such that the path @xmath118 between them does not contain any other critical point . then @xmath118 will be mapped homeomorphically to the path connecting @xmath119 and @xmath120 , which is the complex conjugate of @xmath118 .
since the whole post - critical set consists of critical points , this argument covers the whole hubbard tree .
any non - critical vertices would have to be branch points , and the induced map @xmath110 reverses the cyclic order of edges at branch points , contradicting the fact that @xmath1 is locally conformal at all non - critical vertices .
finally , let @xmath121 and @xmath122 be two edges meeting at a critical vertex @xmath123 .
then @xmath124 , and @xmath125 in @xmath126 , thus @xmath127 .
conversely , if all local degrees are 2 , and all angles are @xmath111 or @xmath112 , complex conjugation @xmath115 satisfies @xmath128 for any two edges meeting at a vertex @xmath94 .
the property that @xmath33 on the tree @xmath13 be expanding is trivially satisfied as there are no julia vertices at all .
@xmath129-symmetric hubbard trees give rise to @xmath129-symmetric polynomials , so this shows the last part of the claim . remark .
the bijective correspondence replaces the vertical edge which intersects the real line by a horizontal edge which is a real interval .
see figure [ figtrees ] for examples of hubbard trees up to degree 8 . proof .
let @xmath116 and @xmath117 be real critical points .
since the hubbard tree is a tree symmetric with respect to the real line , the subtree spanned by @xmath116 and @xmath117 has to be the real interval between them .
since no angle in the hubbard tree is @xmath130 , there can be no other vertex in that interval , i.e. there can not be other critical points between @xmath116 and @xmath117 .
this shows the first claim . in order to construct the bijective correspondence ,
we only need to give the modification of the tree in the upper halfplane .
the part in the lower halfplane is determined by symmetry and the map on the tree is always complex conjugation .
if there are two real critical points @xmath131 , let @xmath132 and @xmath133 be the subtrees in the upper halfplane attached at @xmath116 and @xmath117 , resp .
the correspondence is given by replacing the edge @xmath134 $ ] by @xmath135 $ ] , and attaching @xmath132 and @xmath133 to the left and right of @xmath136 , resp .
, so that @xmath137 .
the inverse of this operation is given by replacing the edge which crosses the real line by a real interval and attaching the left and right subtrees at the endpoints of this interval , again with angles @xmath111 .
observe also that any symmetry with respect to the imaginary axis will be preserved by this correspondence and its inverse .
it remains to count the hubbard trees .
let us start with the easier case of even degree @xmath65 with @xmath139 .
here we have an odd number of critical points , so by claim 2 there is exactly one real critical point @xmath116 .
the number of edges meeting at this point must be even by symmetry . in the case
@xmath140 , there is only one critical point , so there are no edges . in all other
cases the number of edges meeting at @xmath141 must be @xmath113 , since it can be at most 3 by claim 1 . also the angles between the real axis and the edges must be @xmath142 and @xmath111 , respectively .
there are exactly two choices , and they correspond to each other under the symmetry @xmath143 .
since we count hubbard trees up to symmetry , we may fix an orientation .
it also follows that all those hubbard trees have no rotational symmetry .
we can view the part of the tree in the upper halfplane as a finite part of an infinite rooted binary tree . the number of vertices , excluding the root , is @xmath24 , and it is well - known ( and easy to proof from the recursive formula @xmath144 ) that there are @xmath145 of those .
in the case of odd degree @xmath45 there are @xmath146 critical points .
let us count the number of trees which have no vertex on the real line .
then the part in the upper halfplane is again a finite part of a rooted binary tree with @xmath24 vertices , so there are @xmath147 of those .
however , for every such tree the image under @xmath143 has the same form . if @xmath148 , this implies that there are @xmath149 up to symmetry .
however , if @xmath150 , there are trees which are symmetric under @xmath143
. the number of those corresponds to the number of subtrees with @xmath151 vertices attached to the right of the first vertex in the upper halfplane .
this means that there are @xmath152 symmetric trees , so we have @xmath153 inequivalent hubbard trees in this case . by claim 2
, we have the same number of trees with 2 vertices on the real line , so the number of inequivalent hubbard trees is @xmath154 for @xmath155 and @xmath156 for @xmath26 .
let us now count the number of equivalence classes .
the equivalence of two solutions @xmath1 and @xmath12 implies the existence of affine linear maps @xmath15 and @xmath13 with @xmath157 ,
i.e. the existence of an affine map @xmath158 such that @xmath159 maps the critical points of @xmath12 to itself . in particular this implies that @xmath160 is again a solution
. the only rotations @xmath159 which map one of our symmetric hubbard trees to a symmetric hubbard tree are @xmath161 with @xmath162 . in the case
@xmath163 , the map @xmath164 has the same underlying hubbard tree as @xmath12 with the map @xmath165 on it .
this corresponds to the hubbard tree of @xmath12 rotated by 90 degrees with complex conjugation as the map .
this can only happen if @xmath26 , and the number of trees which admit this is @xmath152 .
the case @xmath166 can actually not occur , since the map @xmath164 would have the same hubbard tree and the map on it would be reflection in the line @xmath167 .
however , none of the hubbard trees has one of those lines as a line of symmetry . combining all this ,
the number of equivalence classes is the same as the number of conjugacy classes except in the case @xmath26 , where it becomes @xmath168 .
thus the number of equivalence classes is @xmath169 for all @xmath4 .
9999 bielefeld , b. , fisher , y. and hubbard , j. , the classification of critically preperiodic polynomials as dynamical systems , _ journal of the ams _ , vol . 5 , no .
4 , ( 1992 ) , 721762 .
douady , a. and hubbard , j. , a proof of thurston s topological characterization of rational functions , _ acta .
_ , * 171 * ( 1993 ) , 263297 .
khavinson , d. and neumann , g. , _ on the number of zeros of certain rational harmonic functions _ ,
arxiv : math.cv/0401188 v2 .
khavinson , d. and wiatek , g. , on the number of zeros of certain harmonic polynomials , _ proc .
ams _ , * 131*(2 ) , ( 2002 ) , 409414 . levy , s. , _
critically finite rational maps _ , ph.d . thesis ,
princeton university , 1985 .
pilgrim , k. , canonical thurston obstructions , _ advances in mathematics _ , * 158 * ( 2001 ) , 154168 .
poirier , a. , hubbard forests , ims preprint 92 - 12 , ( 1992 ) .
poirier , a. , on post critically finite polynomials .
part two : hubbard trees , ims preprint 93 - 07 , ( 1993 ) .
rhie , s. h. , _
@xmath4-point gravitational lenses with @xmath170 images _ , arxiv : astro - ph/0305166 v1 . | in @xcite it was proved that harmonic polynomials @xmath0 , where @xmath1 is a holomorphic polynomial of degree @xmath2 , have at most @xmath3 complex zeros .
we show that this bound is sharp for all @xmath4 by proving a conjecture of sarason and crofoot about the existence of certain extremal polynomials @xmath1 .
we also count the number of equivalence classes of these polynomials . |
the electrophoretic motion of charged macroions in an electric field @xmath5 is a longstanding issue in colloid science @xcite .
a complex interplay occurs between hydrodynamic , electrostatic , and thermal forces .
a macroion of charge @xmath4 drags along an electric double layer ( edl ) of small counterions with it and the number of charges `` bound '' in the edl depends on the charge density , surface properties , ion concentration and liphophility of the macroion as well as the specific properties of counterions and salt ions @xcite .
since it is difficult to experimentally probe and control these different parameters at microscopic length and time scales , the interpretation of experimental results often remains unclear .
there is , e.g. , no clear understanding of the variation of the mobility @xmath0 with the surface charge density @xmath1 of the macroion @xcite .
most of the theoretical studies on electrokinetic phenomena consider the coulomb interactions between macroions on the level of the linearized poisson
boltzmann equation @xcite .
this equation , valid for weakly charged macroions and low salt concentration , describes the formation of an edl .
it results in a screened coulomb potential ( yukawa potential ) as the effective interaction potential between the macroions .
the central parameter of this potential is the debye screening length @xmath6 ( with @xmath7 the screening parameter ) that measures the spatial extent of the edl . when a charged colloid moves in the presence of an electric field , the spherical edl is distorted and it is coupled to the hydrodynamic flow of the solvent . in theoretical approaches
@xcite , hydrodynamic effects are incorporated by coupling the poisson equation for the electrostatics to the navier stokes equations . in the resulting electrokinetic equations @xcite , so
called @xmath8 potential is a central quantity .
it is defined as the electrostatic potential @xmath9 at a distance @xmath10 from the colloidal surface where the fluid around the colloid is at rest with respect to the colloid motion .
the imaginary surface at distance @xmath10 from the colloidal surface is called surface of shear or slipping surface @xcite . some of the counterions that are between the surface of the macroion and the surface of shear
are often assumed to move along with the macroion and define the so called _ dynamic _ stern layer @xcite .
the @xmath8 potential can be related to the electrophoretic mobility @xmath11 ( with @xmath12 the steady state macroion velocity ) . in the helmholtz
smoluchowski limit , i.e. for @xmath13 ( with @xmath14 the radius of the macroion ) , the mobility of the macroion can be obtained from the @xmath8 potential via @xmath15 @xcite , where @xmath16 is the permittivity of the surrounding medium and @xmath17 is the shear viscosity of the fluid . in the opposite hckel onsager limit , @xmath18 , the mobility is given by @xmath19 @xcite .
the definition of the potential @xmath20 at the surface of shear is usually based on no slip boundary conditions at the colloidal surface @xcite . since @xmath21 implies a very small debye length , the distance of the surface of shear from the colloidal surface would be typically in the nanometer range . at
such length scales the validity of no slip boundary condition is not clear ( although a hydrodynamic description might be still valid @xcite ) .
calculations of @xmath0 which cover the entire range @xmath22 from the helmholtz smoluchowski limit to the hckel onsager limit have been done by henry @xcite .
but these calculations , valid for low @xmath8 , assumed that the counterion charge density is unaffected by the applied field .
obrien and white @xcite did extensive numerical calculations to calculate @xmath0 versus @xmath8 for different values of @xmath23 in the framework of electrokinetic equations . for @xmath24 , they obtained a pronounced maximum when plotting @xmath0 as a function of @xmath8 .
this behaviour is attributed to the competition between the driving force due to @xmath25 and the retarding forces due to the distortion of the charge cloud constituting the edl .
the driving force increases linearly with @xmath8 ( which is proportional to @xmath4 ) , whereas the retarding force is proportional to @xmath26 .
thus for high values of @xmath8 , @xmath0 decreases with @xmath8 because then the retarding force dominates .
moreover , obrien and white found that for multivalent ions , the position of the maximum shifts to lower values of @xmath8 .
they argue that the distortion of the double layer is increased by multivalent counterions resulting in the increase of retardation forces against the motion of the colloid .
qualitatively similar results were also obtained by the non linear solution by oshina _
et al . _
@xcite and by perturbation expansion in powers of @xmath8 by booth @xcite and overbeek @xcite .
there are also other models of more microscopic origin which look at charged systems , but these models focus mainly on the structure and dynamics of the dynamic stern layer in planar geometry @xcite rather than on electrophoretic properties of charged spherical colloids
. moreover , some of these studies disregard hydrodynamic interactions @xcite or consider only zero temperature properties @xcite .
electrophoresis experiments on colloidal suspensions @xcite allow to determine accurately the electrophoretic mobility @xmath0 as a function of different parameters such as @xmath22 , surface charge density of colloids , etc .
however , it is difficult to disentangle hydrodynamic effects from those due to electrostatic interaction .
moreover , in order to fit experimental data to theoretical models , `` renormalized charges '' of the colloids have to be introduced @xcite . in a recent experimental study by martin
et al . _
@xcite , a maximum was found in the electrophoretic mobility @xmath0 , plotted as a function of the macroion s surface charge density @xmath1 ( note that there is a linear relationship between @xmath1 and the @xmath8 potential in the framework of the linearized poisson
boltzmann equation ) .
the latter maximum occurs in the case of a 2:1 ( counterion - charge : coion - charge ) electrolyte .
a similar result can be also found in the numerical paper by obrien and white @xcite for a 2:1 electrolyte .
however , the experiments were performed for systems of highly charged macroion s . in this case , it is not clear whether the calculations by obrien and white yield a correct description , even on a qualitative level . in this work ,
we aim at addressing this issue by using computer simulation techniques that may elucidate electrophoresis with highly charged particles on a microscopic level .
molecular dynamics ( md ) simulations of colloidal systems suffer from the large separation in length and time scales of solvent and colloidal particles .
the problem of simulating millions of solvent particles to account for hydrodynamic effects can , however , be circumvented by using simulation techniques such as the lattice boltzmann ( lb ) @xcite method or other navier stokes equation solver ( see , e.g. @xcite ) to model the solvent . in this work ,
we employ a hybrid lb / md method @xcite to investigate the electrophoretic motion of a highly charged macroion in an electrolyte solution , thereby modeling macroion and electrolyte in the framework of the so called primitive model ( in a similar way as in ref .
our aim in particular is to investigate those structural and dynamical aspects of the edl which are out of scope of mesoscopic theories . in the following , we are interested in the case of high values of @xmath1 , i.e. the regime beyond the aforementioned maximum in the @xmath27 curve .
in particular the counterion distribution around the macroion is measured to provide insight into the electrical retarding forces affecting the mobility . in order to disentangle hydrodynamic from electrostatic effects ,
the hydrodynamic medium ( lb fluid ) is also switched off , thus simulating the system via langevin dynamics .
systems with monovalent and divalent salt ions are considered . in the latter case , the phenomenon of charge inversion is observed .
the rest of the paper is organized as follows : we describe the lb / md method in sec .
ii . in sec . iiia , we present the results for the monovalent electrolyte solution ( microions ) and in sec .
iiib , we focus on the charge inversion phenomenon with divalent microions . finally , conclusions are given in sec .
the electrophoretic motion of a macroion in an electrolyte solution is studied in the framework of the so called primitive model .
we consider a system of a macroion of charge @xmath28 ( mass @xmath29 a.u . ) and microions of charge @xmath30 ( counterions ) and @xmath31 ( coions ) , each of which are of mass 4 a.u .. the interaction potential between a particle of type @xmath32 and a particle of type @xmath33 ( @xmath34 ) separated by a distance @xmath35 from each other is given by @xmath36 where @xmath37 is the elementary charge and @xmath16 the dielectric constant .
we choose the value @xmath38 ( with @xmath39 the vacuum dielectric constant ) for water at room temperature .
the parameters @xmath40 denote the distance between two ions at contact , @xmath41 , where @xmath42 is the radius of an ion of type @xmath32 . in the following , we use values for @xmath14 that vary from 10 to 20 . for the microions
, we set @xmath43 .
the exponential in eq .
( [ eq_pot ] ) is an approximation to a hard sphere interaction for two ions at contact . for the parameters @xmath44 we choose @xmath45 ev ,
@xmath46 ev , and @xmath47 ev . the parameters @xmath48 are all set to @xmath49 @xmath50 . the long ranged coulomb part of the potential and the forces were computed by ewald sums in which we chose @xmath51 for the constant and a cutoff wavenumber @xmath52 in the fourier part @xcite .
the linear dimension @xmath53 of the simulation box is @xmath54 , using periodic boundary conditions .
all simulations were done at the temperature @xmath55k .
now , the crucial step is to model the solvent . to a first approximation
, the effect of the solvent on colloidal particles ( macroions ) can be described by a langevin equation where one assumes that on the typical time scale of the colloidal particles their collisions with solvent particles are due to gaussian random forces @xmath56 .
these forces lead to a systematic friction force @xmath57 on the colloids , where @xmath58 is the friction coefficient and @xmath59 the velocity of particle @xmath60 at time @xmath61 .
the resulting equations of motion are @xmath62 where @xmath63 and @xmath59 denote respectively position and velocity of a colloidal particle @xmath60 ( @xmath64 with @xmath65 the total number of colloids ) and @xmath66 is the conservative force acting on the particle .
the cartesian components of the random forces , @xmath67 ( @xmath68 ) , are uncorrelated random numbers with zero mean , i.e. @xmath69 the amplitude @xmath70 is determined by the fluctuation dissipation theorem , @xmath71 . in our model ,
the microions are also brownian particles and in thermal equilibrium with the solvent . equations ( [ langevin])([fr2 ] )
still do not provide a complete description of the dynamics in colloidal suspensions . in these equations the mass and momentum transport by the solvent is ignored that leads to the hydrodynamic interactions between the colloidal particles .
however , hydrodynamic interactions can be incorporated rather easily into the langevin description by replacing the absolute velocity @xmath59 of particle @xmath60 in eq .
( [ langevin ] ) by its velocity relative to the fluid velocity field @xmath72 at the position of the particle @xmath63 .
the `` hydrodynamic force '' on particle @xmath60 is then given by @xmath73 + { \bf f}_{{\rm r},i}\ ] ] in a simulation , the velocity field @xmath72 can be calculated from any navier stokes equation solver , whereby thermal fluctuations have to be included in the framework of fluctuating hydrodynamics . in this work
, we use a lattice
boltzmann ( lb ) @xcite scheme to compute @xmath72 .
since the lb method yields the velocity field @xmath74 on a lattice , an interpolation scheme has to be used to determine @xmath74 at the position of the particle @xmath75 .
very recently , ahlrichs and dnweg @xcite have proposed a hybrid md / lb scheme on the basis of the frictional coupling force , eq .
( [ fric ] ) , to simulate polymers in solution . up to now
, we have considered the colloidal particles as point particles . however , real colloids are extended objects with rotational degrees of freedom .
the problem is to `` implant '' an extended object such as a sphere of radius @xmath76 into a lb fluid . to this end , we follow ladd s method @xcite and represent the sphere ( or any other object ) by uniformly distributed boundary points on its surface , where the surface is permeable for the lb fluid .
we have recently proposed a sphere model with 66 boundary nodes of mass @xmath77 with @xmath78 being the total mass of the colloidal particle @xcite .
this model for a spherical particle is also considered in the following .
each of the boundary points is coupled to the lb fluid by the force given by eq .
( [ fric ] ) .
the total force on the particle is determined by the sum over the forces on the boundary nodes ( of course , according to newton s third law , these forces with opposite sign are given back to the lb fluid ) . from this total force
the torque on the particle is calculated , and the total force and torque are then used to update the translational and rotational velocity of the particle , respectively .
more details on our md / lb scheme can be found elsewhere @xcite .
for the lb fluid a standard d3q18 model was used , with which we solved the linearized navier
stokes equations ( the details of the d3q18 model can be found in review articles and the book by succi @xcite ) .
we consider an incompressible fluid in the creeping flow limit ( zero reynolds number ) in our studies .
the kinematic viscosity was set to @xmath79 with @xmath80 the lattice constant and @xmath81 the elementary time unit of the lb fluid .
the density of the lb fluid was set to @xmath82 ( @xmath83 : mass unit of the lb fluid ) .
thermal fluctuations were introduced via the addition of gaussian random numbers to the stress tensor as proposed by ladd @xcite .
the lb fluid is modelled on a cubic lattice with @xmath84 lattice nodes , thus the lattice constant is @xmath85 ( for the linear dimension @xmath54 of the simulation box ) .
the counterions and coions are considered as point particles with respect to the interaction with the lb fluid .
the 66 boundary nodes of the macroion are placed on a sphere of radius @xmath86 . the effective hydrodynamic radius and the viscous retarding force on the macroion
is thus determined by @xmath76 . on the other hand , the variation of @xmath14 from 10 to 20 ( see next section ) allows a change of the macroion s surface charge @xmath1 without changing the hydrodynamic coupling of the particle to the lb fluid . for the friction constant @xmath58 in the coupling force ,
( [ fric ] ) , a total value of @xmath87 is assigned to the macroion which corresponds to @xmath88 for each of the 66 boundary nodes . in a recent publication @xcite , we have shown that this value of @xmath58 corresponds to nearly stick boundary conditions . for the microions ,
a different value for the friction coefficient is used , denoted by @xmath89 in the following .
below we discuss the influence of @xmath89 on the electrophoretic mobility ( see fig .
[ fig6 ] ) . unless otherwise noted , the value @xmath90 was chosen . with respect to the lb fluid ,
the microions are treated as point particles , but we remind the reader that their coulomb radii are set to @xmath43 .
.[tab1 ] charge of the macroion @xmath4 , number of counterions @xmath91 , number of coions @xmath92 , and the corresponding values of the screening parameter @xmath7 and the dimensionless quantity @xmath93 that were used in the simulations for @xmath94 . [ cols="^,^,^,^,^",options="header " , ] we have done md / lb simulations of a single macroion of charge @xmath4 in an electric field @xmath95 pointing in the positive @xmath96 direction .
the charge @xmath4 was varied from @xmath97 to @xmath98 .
the number of counterions and coions ( @xmath91 and @xmath92 , respectively ) used for a given value of @xmath4 are listed in table [ tab1 ] .
also included in this list is the debye screening parameter @xmath99^{1/2}$ ] which was roughly kept constant around 0.13@xmath50 ( the bjerrum length @xmath100 is in our case equal to about @xmath101 ) . for comparison
, we also carried out simulations with a `` langevin dynamics '' ( ld ) where the coupling to the lb fluid was switched off , i.e. @xmath102 .
for the lb / md and ld simulations , the equations of motion were integrated by a heun algorithm with a time step of @xmath103fs .
this very small time step is necessary because we consider explicitly counterions and coions as microscopic particles .
the simulations were done as follows : we first equilibrated the system for 25000 time steps without electric field and without coupling to the lb fluid .
then , the system was coupled to the lb fluid and the electric field was switched on , followed by simulation runs over 400000 time steps .
after 100000 time steps , the steady state was reached and the positions and velocities of the ions were stored every 500 time steps to determine the averaged quantities such as the steady state velocity @xmath12 of the macroion .
now the simulation results are presented for the steady state electrophoretic motion of a macroion in an electrolyte .
systems of highly charged macroions are considered , i.e. their surface charge densities vary between @xmath104@xmath105 ( 38 @xmath0c / cm@xmath105 ) and @xmath107@xmath105 ( 500 @xmath0c / cm@xmath105 ) .
the density of coions ( @xmath108 ) in our simulation box corresponds to a salt concentration of @xmath109mol / l . in this section
, we consider electrolyte solutions that consist of monovalent counter and coions . for a microscopic understanding of electrophoresis , it is of particular interest to display the _ dynamic _ distribution of counterions in the vicinity of the macroion , i.e. in the edl .
this distribution is determined by an interplay between the electrostatic attraction and the hydrodynamic flow around the macroion .
some of the counterions in the edl move along the same direction as the macroion , whereas , due to the electric field , other counterions in the edl are accelerated in the opposite direction .
we analyze the dynamic behavior of the counterions by their average velocity @xmath110 as a function of the radial distance @xmath35 from the center of the macroion and the cumulative counterion charge @xmath111 around the macroion .
a ) @xmath112 as a function of @xmath113 for different values of the electric field , as indicated .
data is shown for 1 macroion of charge @xmath114 , @xmath115 monovalent counterions and @xmath116 monovalent coions .
b ) cumulative counterion charge @xmath111 for the same parameters as in a ) .
@xmath2 is the cumulative counterion charge at which @xmath110 is zero .
the determination of @xmath2 is indicated in the figure by the dashed lines for the example @xmath117 v / . ] first , we consider a system consisting of one macroion of charge @xmath118 , @xmath115 monovalent counterions and @xmath108 monovalent coions .
for this case , fig .
[ fig1 ] displays @xmath119 ( with @xmath12 the steady state macroion velocity ) and @xmath111 for different choices of the electric field @xmath25 . within the statistical errors ,
the results essentially coincide for the two lowest values of the electric field , @xmath117v / and @xmath120v / . the choice @xmath117v / is used for all further results presented in this work , in particular for the estimates of the electrophoretic mobility , @xmath121 .
as shown previously @xcite , for @xmath122 the linear response regime is essentially achieved for @xmath123v / . from the behavior of @xmath112 ,
different regions can be identified with respect to the distance from the center of the macroion .
close to the macroion s surface a layer with a thickness of about @xmath124 is formed where @xmath125 is constant with a value around @xmath126 for the two lowest values of @xmath25 . in the following , this region is called the dynamic stern layer , where , due to the electrostatic attraction by the macroion , counterions are essentially condensed onto the surface of the macroion . beyond the stern layer , the counterion velocity changes quickly its sign , thus indicating a motion in the direction opposite to that of the macroion . the point where the counterion velocity vanishes can be used as a measure of the extent of the stern layer . as shown in fig .
[ fig1 ] , the cumulative counterion charge at this point , @xmath2 , is significantly smaller than the bare charge of the macroion ( e.g. @xmath127 for @xmath117v / ) . not until @xmath35 is of the order of 34@xmath14 , the counterion charge @xmath111 is equal to the macroion s charge , thus completely neutralizing the latter . for high values of @xmath25 , e.g. @xmath128v / , the counterions are stripped off the surface of the macroion , since the force due to the electric field dominates over the coulomb attraction between macroion and counterions .
this leads to a lower value of @xmath2 for high values of @xmath25 and a less efficient shielding of the macroion compared to the case of low values of @xmath25 .
this in turn increases the velocity @xmath12 of the macroion and thus explains the low values of @xmath125 for large values of @xmath25 at distances far from the colloidal surface .
@xmath112 as a function of @xmath113 for two different values of the bare charge @xmath4 , as indicated . for the hd case , the data from fig . [ fig1 ] for @xmath129e and
@xmath117v / are plotted ( open circles ) .
the closed circles are the corresponding results from ld simulations . also shown are results for @xmath130 ( open and closed diamonds for hd and ld , respectively ) . in this case
, the system contains @xmath131 counterions and @xmath132 coions . ]
one might expect , that the region between the stern layer and the point where @xmath133 , is strongly affected by hydrodynamic flow features . that this is indeed the case
can be infered from a comparison to ld simulations where the coupling to the lb fluid is switched off ( in the following , we refer to simulations with a coupling to lb as hd simulations ) .
the macroion mobility @xmath0 as a function of the surface charge density @xmath1 for ld and hd simulations as indicated .
the charge density @xmath1 is either varied by changing the radius @xmath14 of the macroion from @xmath134to @xmath135 keeping @xmath4 fixed at @xmath136 or @xmath137 , or by changing @xmath4 from @xmath138 to @xmath139 keeping @xmath14 fixed at @xmath135 .
the number of counterions and coions used for each value of @xmath4 is mentioned in text . for all data ,
the `` hydrodynamic radius '' is chosen to be constant at @xmath140 . ] in fig .
[ fig2 ] , @xmath112 from ld simulations is compared to the same quantity from hd simulations for the two macroion charges @xmath136 and @xmath139 ( in the latter case the system contains @xmath141 counterions and @xmath132 coions ) . in the hd case
, we see that for @xmath139 , the ratio @xmath112 is very close to one in the stern layer region .
however , for @xmath136 the ratio is significantly smaller .
the strong coulomb attraction dominates the viscous drag and thermal fluctuations for high charges , such that layers of counterions nearest to the colloidal surface nearly stick to the surface .
but for lower macroion charges , one starts to see deviations from the assumption that the _ dynamic _ stern layer consists of immobile counterions @xcite .
moreover , the stern layer is extended , and it is not restricted to one layer of microions closest to the macroion surface as observed in @xcite .
this is due to the much higher values of the surface charge density @xmath1 used in our study . as indicated by fig .
[ fig2 ] , in the ld case the motion of the counterions is less correlated to that of the macroion than in the hd case .
the value of @xmath112 is smaller in the stern layer region than in the corresponding hd data .
moreover , the counterions reverse their velocity at about @xmath142 away from the macroion surface , followed by a more rapid decrease of @xmath112 than in hd . from data presented below ( in fig . [ fig4 ] ) , we will see that the number of counterions carried along in the stern layer is almost the same for ld and hd , which reveals that , in the ld case , the counterions are more densely packed in the stern layer . in order to investigate the mobility @xmath0 of the macroion as a function of its surface charge density @xmath143
, two different types of simulations were performed .
first , the charge @xmath4 was varied from @xmath138 to @xmath98 while keeping the radii @xmath144 and @xmath14 fixed at @xmath134 and @xmath135 , respectively .
the number of counter and coions used for a given value of @xmath4 are listed in table [ tab1 ] . also included in this list
is the debye screening parameter @xmath145^{1/2}$ ] which was roughly kept constant around @xmath146@xmath50 .
secondly , runs for @xmath136 and @xmath137 were done in which the charge density @xmath1 was varied by choosing different macroion radii 10@xmath147 20 in steps of @xmath148 or @xmath149 ( note that the radius @xmath76 remains fixed at 10 in these runs ) .
the results for the different runs are shown in fig .
[ fig3 ] where @xmath0 is plotted as a function of @xmath1 for both hd and ld runs . in both cases , the mobility @xmath0 decreases with increasing @xmath1 .
this behavior is in qualitative agreement with theoretical calculations for high values of the surface charge , i.e. @xmath150 ( see ref .
@xcite and references therein ) .
note that , at small @xmath1 the opposite behavior is observed , i.e. an increase of @xmath0 with @xmath1 @xcite .
ratio @xmath3 as a function of @xmath1 .
as indicated , in the different data sets either @xmath4 or @xmath14 are fixed while @xmath1 is varied ( see also fig . [ fig3 ] ) . ] for monovalent microions , fig .
[ fig3 ] demonstrates that at constant screening parameter @xmath7 the mobility @xmath0 is controlled by the surface charge density of the macroion .
discrepancies that can be seen in the plot might stem from the slight variation of @xmath7 in the different runs ( see table [ tab1 ] ) .
the ld runs that are shown in the figure exhibit a similar qualitative behavior .
however , compared to hd , the ld data is shifted towards lower values of @xmath0 .
this is because the hydrodynamic flow field is switched off in the ld simulations .
the coupling of the lb fluid to the motion of the macroion leads to a backflow effect in the lb fluid which enhances the mobility of the macroion .
this backflow effect yields also stronger correlations between the motion of the counterions and that of the macroion .
thus , as shown in fig .
[ fig2 ] , the reduced counterion velocity @xmath119 decays slower in the hd case than in the ld case . the correlations between counterions and macroions are further considered in fig .
[ fig4 ] . here , @xmath3 is plotted vs. @xmath1 .
as before , we vary @xmath1 by changing @xmath14 keeping @xmath4 fixed or by changing @xmath4 with @xmath14 fixed at @xmath135 . as we see , the different data sets fall roughly onto one master curve ( note that this is even true for the ld data ) .
the functional behavior of @xmath3 reflects the one for @xmath0 . at small values of @xmath1 , the ratio @xmath3
is `` rapidly '' increasing ( associated with a rapid decrease of @xmath0 ) and it seems to saturate at high @xmath1 ( as @xmath0 does ) .
thus , the electrophoretic mobility shows a saturation when @xmath3 is approaching unity . in the latter case
, the electric field sees a `` particle '' with an effective charge @xmath151 .
plot of the measure of distortion @xmath152 of the stern layer , normalised by the coulomb radius of colloid @xmath14 as a function of the inverse surface charge density @xmath153 ( for the definition of ` @xmath152 ' see text ) .
the inset shows @xmath154 which is a measure of the charge separation due to the distortion of the edl . ]
the regime of low macroion charges @xmath4 has been studied recently by lobaskin _
et al . _
@xcite using a similar lb / md technique .
these authors considered a primitive model in the salt free regime ( i.e. without coions ) .
their results agree with experimental data @xcite for low @xmath4 , where an increase of @xmath0 with @xmath4 ( @xmath1 ) is observed . in this regime , @xmath0 is controlled by the bare charge @xmath4 as there is no significant dynamic stern layer to create a shielding for @xmath4 .
hence the peak in @xmath0 observed in the experimental data of martin - molina _
et al . _
@xcite denotes the point of crossover from this regime to a regime where @xmath0 is controlled by @xmath3 . for high values of @xmath1
, @xmath0 decreases with @xmath1 due to greater shielding for higher @xmath4 , whereas , for low values of @xmath1 ( @xmath4 ) the mobility increases with @xmath1 .
obrien and white @xcite have proposed that the decrease in macroion mobility can be attributed to the increasing distortion of the edl at high values of @xmath1 . to see whether the distortion of the charge cloud is indeed a relevant effect also at very high values of @xmath1 ( considered in this work ) , we quantify the distortion of the edl as follows : we compute the distance @xmath152 of the center of mass ( cm ) of the edl from the center of the macroion . here , those counterions form the edl which are within the _ dynamic _ stern layer and move along the macroion .
figure [ fig4a ] shows that the normalised distortion @xmath155 and the normalised quantity @xmath154 decrease with increasing @xmath1 .
the quantity @xmath154 is a measure of the charge separation between the center of the macroion and center of the counterion cloud .
thus , the amount of distortion becomes less pronounced with increasing @xmath1 .
this rules out the mechanism proposed by obrien and white in the case of high values of @xmath1 .
plot of @xmath156 versus the charge density @xmath1 .
the mobilities @xmath157 and @xmath158 correspond to hd and ld runs , respectively . ]
we have already infered from fig .
[ fig2 ] that hydrodynamic backflow due to the lb fluid enhances the correlations between the counterion and the macroion motion as well as the absolute value of @xmath0 . in fig .
[ fig5 ] , the mobilities obtained from ld and hd runs ( denoted by @xmath157 and @xmath158 , respectively ) are directly compared to each other by plotting the ratio @xmath156 as a function of @xmath1 .
we see that @xmath156 increases almost linearly with @xmath1 .
macroion mobility @xmath0 versus the counterion coupling constant @xmath89 for ld and hd simulations .
the colloidal friction @xmath58 is fixed at @xmath159 and @xmath160 .
the lines are guides to the eye . ]
microscopically , we understand this by noting that the higher macroion charge density leads to a higher number density of counterions per unit volume in the moving edl of the colloid .
the counterion coupling to the lb fluid results in an increased velocity of the lb fluid around the macroion at higher densities .
thus the macroion feels less frictional drag and , compared to the ld case , a higher value of @xmath0 is obtained . in fig .
[ fig6 ] , it is further explored how the counterion coupling with the lb fluid affects the colloidal mobility .
we keep the friction coefficient for the macroion coupling to the lb fluid , @xmath58 , fixed and vary that for the microion coupling , @xmath89 ( for the definition see sec .
ii ) . whereas , for the hd simulations
, @xmath0 decreases significantly with @xmath89 , in the ld case , @xmath0 exhibits only a weak dependence on @xmath89 .
thereby , the hd result tends to approach the one from ld for high values of @xmath89 . in the hd case ,
small values of @xmath89 mean a weaker coupling of the microion motion to that of the lb fluid , leading to an increase of the macroion mobility .
the detailed understanding of this finding requires further investigation .
note that , in the ld simulation for @xmath138 , a dynamic stern layer can not be formed , because the coulomb attraction is too weak in this case . hence , a high value of mobility results which is outside the range used in fig . [ fig6 ] . on the other hand , in the hd simulation for @xmath138 a well - defined dynamic stern layer is seen and , as shown by fig . [ fig6 ] ,
the behavior of @xmath0 is qualitatively similar to that at @xmath136 .
this demonstrates the importance of the hydrodynamic flow field for the behavior of the electrophoretic mobility of weakly charged colloids .
electrophoretic mobility @xmath0 as a function of surface charge density @xmath1 for two different systems using ld and hd , as indicated .
the lines serve as guides to the eye . ]
cumulative charge of divalent ions @xmath111 as a function of @xmath161 for different values of macroion radius @xmath14 as indicated .
data for the system with @xmath162 divalent counterions and @xmath163 divalent coions is shown ( the charge of the macroion is @xmath136 ) . the distance @xmath35 in the x - axis is normalised by @xmath164 which corresponds to the largest radius used . ]
charge inversion in the presence of multivalent counterions has been observed in static experiments and theoretical studies @xcite .
experimentally charge inversion is detected using mainly electrophoresis .
it is well established that charge inversion is due to counterion correlations and beyond the scope of mean field theories .
our aim in this study is twofold : the onset charge density @xmath1 , at which charge inversion occurs , is determined and secondly , the extent of the stern layer from hd simulations is compared to that from ld simulations . to this end ,
two systems of a macroion of charge @xmath136 mixed with divalent microions are considered , one with @xmath162 divalent counterions and @xmath163 divalent coions , and the other one with @xmath165 divalent counterions and @xmath166 divalent coions . in each case , 3 monovalent counterions are added to maintain charge neutrality .
the charge density @xmath1 is varied by changing the radius of the macroion @xmath14 from @xmath134 to @xmath135 , thereby keeping the radius @xmath76 fixed at @xmath134 . in fig .
[ fig7 ] , the mobility @xmath0 as function of @xmath1 is plotted for the different systems .
different from the case of monovalent microions , @xmath0 changes its sign in all cases , thus indicating that at high values of @xmath1 the macroion is moving opposite to the direction of the electric field .
overall , at a given value of @xmath1 , the mobility of the macroion in the presence of divalent microions is smaller than with monovalent ones .
since the screening due to divalent microions is much more effective , the counterion charge density is higher near the colloidal surface than in electrolytes with monovalent microions .
this can be seen in fig .
[ fig8 ] , where the system with @xmath163 coions is considered . the cumulative charge @xmath111 around the macroion is displayed for different values of @xmath14 corresponding to different values of @xmath1 at the fixed macroion charge @xmath136 .
we see that for @xmath167 , @xmath111 develops a local minimum near the colloidal surface .
this indicates charge inversion ( i.e. @xmath168 ) and it explains the appearance of negative values of @xmath0 . for comparison , we have plotted ld data in fig .
[ fig8 ] for the largest and the smallest radii used , i.e. , for @xmath94 and @xmath169 , respectively .
we see that the charge profile for ld and hd is nearly identical , which shows that the dynamic stern layer is determined by the static coulomb forces . for higher salt concentration ,
charge inversion is more pronounced , as seen by lower values of @xmath0 at high @xmath1 .
both ld and hd show the phenomenon of charge inversion at high values of @xmath1 .
in particular in the hd case , charge inversion is more pronounced for higher salt concentration . as in the studies with monovalent counter and coions , for divalent microions @xmath170 holds as well .
in this work , a hybrid md / lb scheme was used to investigate the electrophoretic mobility of a macroion in an electrolyte solution .
we have considered highly charged macroions ( @xmath171c / cm@xmath105@xmath172c / cm@xmath105 ) for which theories based on the linearized electrokinetic equation are not applicable . for high values of the macroion s surface charge @xmath1 ,
experiments on systems with multivalent salt ions @xcite have shown that the electrophoretic mobility @xmath0 can decrease with increasing @xmath1 for @xmath173 salts whereas @xmath0 for @xmath174 salts , it shows a plateau . in our simulations , we observe decrease of @xmath0 with @xmath1 for both @xmath175 and @xmath174 salt . moreover , in contrast to previous theoretical studies @xcite , we observe a lowering of @xmath0 with @xmath1 even for @xmath176 for monovalent salt ions .
the authors of @xcite , attribute the lowering of @xmath0 in the @xmath173 case to the presence of coions in the edl . as we have demonstrated in this work , a decreasing @xmath0 as a function of @xmath1 requires the formation of a pronounced dynamic stern layer of charge @xmath2 , consisting of counterions that move along the same direction as the macroion in the presence of an electric field .
the lowering of @xmath0 is directly related to the increase of the quantity @xmath3 which we call the screening charge fraction .
the absolute value of screening charge fraction @xmath177 of the colloid
counterion cloud complex increases with increasing @xmath1 , reflecting a lowering of @xmath0 .
note that the number of coions in the dynamic stern layer is negligible .
furthermore , we observe charge inversion for divalent counter and coions , but only when we reach sufficiently high values of @xmath1 . at very high values of @xmath1 , @xmath0 for colloidal systems with monovalent salt ions
tends to saturate , because there is no space to add more counterions into the dynamic stern layer .
however , this is different when multivalent salt ions and counterions are used .
then , @xmath178 holds and @xmath0 becomes negative .
the distribution of charge in the edl for this case as a function of distance @xmath35 from the center of the colloid is non monotic and completely different from what is expected from standard electrokinetic theories @xcite . though we have not carried out studies for the 2:1 salt case , it is possible that the discrepancy in the experimental results between the 2:1 and the 2:2 salt case arises from entropic / osmotic forces . the larger number of coions in the 2:1 salt ( compared to the systems with a 2:2 salt ) may lead to larger osmotic forces on the counterions which could result in larger values of @xmath2 for the 2:1 case .
however , this issue can be resolved only after further investigations . in the work of obrien and white @xcite
, the lowering of @xmath0 with the @xmath8 potential has been attributed to the distortion of the edl .
though the explicit measurement of the retarding force arising from the distortion of the edl is outside the scope of this work , we have shown that the charge distortion within the dynamic stern layer decreases with increasing @xmath1 .
this is indicated by the lowering of the quantities @xmath155 and @xmath154 with increasing @xmath1 .
this observation combined with a host of other evidences presented in this paper , indicate that accumulation of charges in the stern layer is the dominant mechanism for the lowering of @xmath0 as a function of @xmath1 , providing that high values of @xmath1 are considered . by comparing the lb / md ( or hd ) results to those from ld simulations , we were able to elucidate the role of hydrodynamic interactions . the structure of the stern layer is more compact in the langevin simulations .
hydrodynamic interactions enhance the electrophoretic mobility and so they have the opposite effect to electrostatic screening .
this is due to a backflow effect in the fluid which pushes the counterions to move along the same direction as the macroion .
hydrodynamic forces have pronounced influence in the formation of the dynamic stern layer , especially for low charges .
an interesting finding is that the mobility as obtained from hd relative to that from ld exhibits an almost linear increase with @xmath1 .
further work on the role of hydrodynamic interactions for the electrophoresis of charged colloids is in progress .
* acknowledgment : * we thank burkhard dnweg , vladimir lobaskin , and thomas palberg for stimulating discussions .
we acknowledge financial support by the deutsche forschungsgemeinschaft ( dfg ) under grants no .
ho 2231/1 and ho 2231/2 , by the sfb 625 `` von einzelnen moleklen zu nanoskopisch strukturierten materialien '' , and by the sfb tr6 `` colloidal dispersions in external fields '' .
generous grants of computing time on the jump at the nic jlich are gratefully acknowledged .
99 j. lyklema , _ fundamentals of colloid and interface science _
( academic , new york , 1995 ) .
j. lyklema and m. minor , coll .
a * 140 * , 33 ( 1998 ) .
r. j. hunter , _ foundations of colloid science _ ( oxford university press , oxford , 2001 ) .
m. lozada
cassou , e. gonzalez tovar , and w. olivares , phys . rev .
e * 60 * , r17 ( 1999 ) .
russel , d.a .
saville , and w.r .
schowalter , _ colloidal dispersions _ ( cambridge university press , cambridge , 1989 ) .
p. h. wiersema , a. l. loeb , and j. t. g. overbeek , j. coll .
sci . * 22 * , 78 ( 1966 ) ; j. th .
t overbeek , kolloidchem .
beihefte . ,
* 54 * , 287 ( 1943 ) .
tata , curr . sc . * 80 * , 948 ( 2001 ) .
k. besteman , m.a.g .
zevenbergen , h.a .
heering , and s. g. lemay , phys .
. lett . * 93 * , 170802 ( 2004 ) .
a. martin
molina , m. quesada
perez , f. galisto gonzalez , and r. hidalgo alvarez , j. phys .
b * 106 * , 6881 ( 2002 ) .
v. lobaskin , b. dnweg , and c. holm , j. phys . : condens .
matter , * 16 * s4063 ( 2004 ) .
v. lobaskin , m. medebach , th .
palberg , b. dnweg , and ch .
holm , preprint cond - mat/0601588 .
torrie and j.p .
valleau , j. chem . phys . * 73 * , 5807 ( 1981 ) . r.r .
netz , phys .
lett . * 91 * , 138101 ( 2003 ) .
l. joly , ch .
ybert , e. trizac , and l. bocquet , phys .
lett . * 93 * , 257805 ( 2004 ) .
j. horbach and s. succi , phys .
. lett . * 96 * , 224503 ( 2006 ) .
m. patra , m. patriarca and m. kartunen , phys .
e * 67 * , 031402 ( 2003 ) .
p. wette , h.j .
schpe , and t. palberg , j. chem . phys . *
116 * , 10981 ( 2002 ) .
m. medebach and t. palberg , j. chem . phys . *
119 * , 3360 ( 2003 ) .
m. medebach and t. palberg , j. phys . : condens .
matter * 16 * , 5653 ( 2004 ) .
t. palberg , m. medebach , n. garbow , m. evers , a. bareira fontecha , and h. reiber , j. phys . : condens .
matter * 16 * , s4039 ( 2004 ) .
l. shapran , m. medebach , p. wette , t. palberg , h. j. schpe , j. horbach , t. kreer , a. chatterji , coll .
a * 270271 * , 220 ( 2005 ) .
grosberg , t.t .
nguyen , and b.i .
shklovskii , rev .
phys . * 74 * , 329 ( 2002 ) .
r. benzi , s. succi , and m. vergassola , phys . rep .
* 222 * , 145 ( 1992 ) ; s. chen and g. doolen , ann .
fluid mech .
* 130 * , 329 ( 1998 ) ; s. succi , _ the lattice boltzmann equation _
( oxford univ . press , oxford , 2001 ) .
k. kim , y. nakayama , and r. yamamoto , phys .
lett . * 96 * , 239902 ( 2006 ) .
a. chatterji , j. horbach , j. chem . phys . * 122 * , 184903 ( 2005 ) .
a. chatterji and j. horbach , mathematic and computers in simulation . ,
accepted for publication ( 2006 ) .
t. kreer , j. horbach , a. chatterji , _ nonlinear effects in charge stabilized colloidal suspensions _
e , accepted for publication ( 2006 ) ; preprint cond - mat/0602250 .
m.p . allen and d.j .
tildesley , _ computer simulation of liquids _
( clarendon press , oxford , 1987 ) .
d. frenkel and b. smit , _ understanding molecular simulation : from algorithms to applications _ , 2nd edn ( academic press , san diego , 2002 ) .
a.j.c ladd , j. fluid mech .
* 271 * , 285 ( 1994 ) ; j. fluid mech . * 271 * , 311 ( 1994 ) .
p. ahlrichs and b. dnweg , j. chem . phys . *
111 * , 8225 ( 1999 ) .
r. messina , c. holm , and k. kremer , phys .
* 85 * , 872 ( 2000 ) ; r. messina , c. holm , and k. kremer , europhys .
lett . * 51 * , 461 ( 2000 ) . | using computer simulations , the electrophoretic motion of a positively charged colloid ( macroion ) in an electrolyte solution is studied in the framework of the primitive model . in this model ,
the electrolyte is considered as a system of negatively and positively charged microions ( counterions and coions , respectively ) that are immersed into a structureless medium .
hydrodynamic interactions are fully taken into account by applying a hybrid simulation scheme , where the charged ions ( i.e. macroion and electrolyte ) , propagated via molecular dynamics ( md ) , are coupled to a lattice boltzmann ( lb ) fluid . in a recent electrophoretic experiment by martin - molina _
et al . _ [ j. phys .
chem .
b * 106 * , 6881 ( 2002 ) ] , it was shown that , for multivalent salt ions , the mobility @xmath0 initially increases with charge density @xmath1 , reaches a maximum and then decreases with further increase of @xmath1 .
the aim of the present work is to elucidate the behaviour of @xmath0 at high values of @xmath1 . even for the case of monovalent microions
, we find a decrease of @xmath0 with @xmath1 .
a dynamic stern layer is defined that includes all the counterions that move with the macroion while subject to an external electrical field .
the number of counterions in the stern layer , @xmath2 , is a crucial parameter for the behavior of @xmath0 at high values of @xmath1 . in this case
, the mobility @xmath0 depends primarily on the ratio @xmath3 ( with @xmath4 the valency of the macroion ) . the previous contention that the increase in the distortion of the electric double layer ( edl ) with increasing @xmath1 leads to the lowering of @xmath0 does not hold for high @xmath1 .
in fact , we show that the deformation of the edl decreases with increase of @xmath1 . the role of hydrodynamic interactions is inferred from direct comparisons to langevin simulations where the coupling to the lb fluid is switched off .
moreover , systems with divalent counterions are considered . in this case , at high values of @xmath1 the phenomenon of charge inversion is found . |
radiative transfer effects are capable of producing line spectra that differ substantially from those emitted under optically thin conditions . as resonance line photons travel through a medium that is optically thick to its own radiation , multiple non - coherent scattering substantially alters the line profiles of the emergent spectrum .
line transfer also affects the emergent line intensities through line or continuum absorption , in which case the energy is either ( 1 ) re - radiated through other discrete transitions or ( 2 ) is used in heating the ambient gas . in hydrogen - like ions , for example , the upper levels of the higher - series lyman transitions ( @xmath5 , @xmath6 , @xmath7 , etc . )
have finite probabilities of decaying to excited states .
this implies that the lyman line photons will be destroyed and re - radiated in the balmer , paschen , etc .
series if the line optical depths are sufficiently large .
ratios of lyman to balmer line intensities , for example , can then be used to estimate the optical depth , as well as other local physical parameters , such as the density and temperature ( see , e.g. , @xcite , and references therein ) .
another well - known example of such an effect is the bowen fluorescence mechanism @xcite , where it was realized that the coincidence in the wavelengths of ( he@xmath8 ) ly@xmath0 and ( o@xmath9 ) @xmath10 @xmath11 transitions allows the conversion of ly@xmath0 line photons into the upper levels of , which then decay to produce lines in the optical / uv band known as the bowen lines at @xmath12 .
these line were commonly found to be extraordinarily strong in the optical spectra of planetary nebulae and seyfert galaxies .
@xcite presented the first complete analysis of the ly@xmath0 transfer and the conversion efficiencies of the bowen lines , and demonstrated that the observed spectral properties can be reasonably well - described by this mechanism .
subsequently , @xcite presented a more detailed calculation using an exact form of the redistribution function and a realistic ionization balance calculation ( see , also , @xcite ) . in the simplest approximation generally adopted for spectral modeling
, one assumes that line photons created in the gas escape the medium without any further interaction with the constituent atoms in the medium .
this is generally referred to as the optically - thin limit or the coronal approximation in the context of collisionally ionized plasma . in this limit
, the ionization balance is determined solely by the local temperature and is assumed to be entirely decoupled from the radiation field .
the emission spectra computed under these assumptions seem to apply fairly well to the observed x - ray spectra of a wide variety of sources including stellar coronae , supernova remnants , hot diffuse gas in starburst galaxies , and the intergalactic medium in clusters of galaxies . for photoionized plasmas , on the other hand , codes such as cloudy @xcite and xstar @xcite treat the transfer of x - ray lines using either the escape probability method or the assumption of complete redistribution . while these assumptions yield reasonable results for _ isolated _ lines with moderate line optical depths ( @xmath13 )
, it does not properly describe line scattering when absorption in the damping wings becomes important ( i.e. , for @xmath14 depending on the voigt parameter ; see , e.g. , @xcite ) . such conditions might be relevant for a wide variety of astrophysical environment including ionized surface layers of accretion disks , extended circumsource regions in agn , and possibly shocked regions in the stellar wind of hot young stars .
the purpose of this paper is to demonstrate that radiative transfer effects can alter intensities of some of the brightest emission lines in the x - ray band , in particular , the hydrogen- and helium - like lines from carbon , nitrogen , and oxygen .
non - solar cno abundance ratios have been inferred in a number of sources observed with the grating spectrometers on _ chandra _ and _ xmm - newton_. the relativistic emission line interpretation of the _ xmm - newton _ rgs spectra of the seyfert 1 galaxies mcg@xmath156 - 30 - 15 and mrk 766 presented by @xcite , for example , requires the nitrogen emission lines , which are presumably formed in an optically thick accretion disk , to be much stronger than that of the oxygen line assuming solar abundance ratios ( see also , @xcite ) .
similarly , @xcite conclude that nitrogen in the extended circumnuclear regions in the seyfert 2 galaxy ngc 1068 is overabundant by a factor of @xmath16 based on measurements of the hydrogen- and helium - like line intensities with the rgs ( see also , @xcite ) .
@xcite also report abnormally high nitrogen line intensities relative to those of oxygen in the rgs spectrum of the low - mass x - ray binary her x-1 , and interpreted as an overabundance of nitrogen due to h - burning by the cno - cycle .
all of the modeling , however , are based on rather simple models that do not include detailed line transfer .
we show that the effects of line overlap alone can explain some of the observed anomalies , in particular the strength of the nitrogen lines , if the optical depth is large , and that detailed transfer models are required to infer accurate values for the cno abundances with x - ray spectroscopic data .
we note , however , that the theoretical model presented here is highly idealized in order to simply illustrate the general effects and describe the relevant spectroscopic details , and is certainly not adequate for quantitative comparisons with observations of complicated systems .
the paper is organized as follows . in 2 , we describe the spectroscopic properties of the frequency range near the ( o@xmath17 ) ly@xmath0 doublet transitions , which contains a pair of overlapping ( n@xmath18 ) ly@xmath3 lines and underlying sources of continuum opacity , which is usually dominated by k - shell absorption of h- and he - like carbon ( c@xmath19 and c@xmath20 ) and he - like nitrogen ( n@xmath18 ) . we compute the destruction probabilities of ly@xmath0 due to line absorption by the ly@xmath3 transition and continuum absorption by carbon and nitrogen for temperatures ranging from cool photoionized media ( @xmath21 ) to hot collisionally ionized plasmas ( @xmath22 ) , which is described in 3 .
we demonstrate that the wavelength coincidence allows efficient conversion of ly@xmath0 line radiation into the lines , and may be misidentified as a cno abundance anomaly ( 4 ) .
finally , in 5 , a few other potentially important overlaps are identified and their effects on the global x - ray spectrum under optically thick conditions are briefly discussed .
in the non - relativistic , hydrogenic approximation , the transition energy of the ly@xmath0 line ( @xmath23 ) coincides exactly with that of the ly@xmath3 ( @xmath24 ) line ; i.e. , @xmath25 holds exactly for @xmath26 , @xmath27 and @xmath28 , @xmath29 .
more precisely , the ly@xmath0 transition is a doublet at wavelengths @xmath30 ( @xmath31 ) and @xmath32 ( @xmath33 ) that correspond to decays to the ground state from the @xmath34 and @xmath35 levels , respectively .
the oscillator strength of the ly@xmath36 line is twice that of the ly@xmath37 line .
the ly@xmath3 transition is also a doublet with @xmath38 ( @xmath39 ) and @xmath40 ( @xmath41 ) .
since the wavelength difference of @xmath42 m is much too small to be resolved at any reasonable temperature where n@xmath18 can exist , we simply treat them as a single line with an oscillator - strength - averaged wavelength of @xmath43 ( @xmath44 ; see , figure [ fig : grot ] ) .
the ly@xmath0 and ly@xmath3 transition wavelengths are adopted from @xcite and @xcite , respectively , which include reduced - mass and quantum electrodynamical corrections to the eigenvalues of the dirac equation .
oscillator strengths and radiative decay rates are calculated using the flexible atomic code ( fac ; @xcite ) .
all of the relevant atomic parameters are listed in table [ tbl1 ] .
the relative cross sections assuming solar abundances ( @xmath45 = 0.13 ; @xcite ) and a temperature of @xmath46 , typical for a photoionization - dominated medium at this level of ionization , are shown in figures [ fig : cross10 ] .
at this temperature , the ly@xmath3 lies approximately 2 doppler widths away from the ly@xmath37 line and 10 doppler widths from the ly@xmath36 line .
the cross section is almost three orders of magnitude lower than that of the lines , which implies that significant scattering between the lines occur when the total optical depth is @xmath47 .
similarly , at a temperature of @xmath48 , the ly@xmath3 lies @xmath49 doppler width away from the ly@xmath37 line and 4 doppler widths from the ly@xmath36 line , as shown in figure [ fig : cross50 ] .
the emission coefficient in this spectral range is dominated almost entirely by the ly@xmath0 lines .
the intrinsic ly@xmath3 emission coefficient is approximately two to three orders of magnitude lower in both collision - dominated and recombination - dominated plasmas with solar @xmath45 abundance ratio .
if the medium is photoionized by some external source of continuum radiation , photoexcitation may enhance the line emissivities of the lines , but this is important only in the surface layer within @xmath13 , as external radiation will not be able to penetrate deeper into the medium .
since we are primarily interested in the properties of the bowen mechanism at large optical depths , we ignore the ly@xmath3 intrinsic source term entirely .
absorption of an line photon by is followed by one of the following two processes : ( 1 ) re - emission in the ly@xmath3 line , which results in a pure scattering event or ( 2 ) destruction of the line through cascades via the upper levels ( @xmath50 ) , followed eventually by decay to the ground level .
collisional excitation and de - excitation from these levels are important at only extremely high electron densities ( @xmath51 ) , and so we do not consider these processes here .
we note , however , that thermalization of the line photons via collisions may occur at sufficiently high density and high optical depth , such that @xmath52 , where @xmath53 is the mean number of scatterings experienced by the line photons as they propagate through the medium .
the branching ratios for the two possible processes can be determined from the radiative decay rates from the @xmath54 level to all possible levels .
we use radiative rates calculated with fac except for the two - photon decay rates , which are adopted from @xcite .
the branching ratios for processes ( 1 ) and ( 2 ) are found to be 0.797 and 0.203 , respectively .
the final fate of the 20.3% of the line photons that initially decay to an upper level can be understood by computing the cascade matrices ; quantities that represent the probabilities of one state decaying to another state via all possible channels @xcite .
we find , for example , that 58% of them produce the balmer , paschen , etc .
lines and eventually result in two - photon emission from the @xmath55 level , while the remaining 42% produce the lyman lines in the soft x - ray band . continuum opacity from ions that coexist with is dominated primarily by ( ionization potential @xmath56 ) , ( @xmath57 ) , and ( @xmath58 ) for both collisionally and photoionized plasmas in ionization and thermal equilibrium . the threshold wavelength of ( @xmath59 ) lies shortward of the ly@xmath0 line .
lower charge states of carbon and nitrogen may contibute very small amounts of continuum opacity as well . for solar abundances , the typical value for the combined continuum opacity relative to the ly@xmath0 line opacity is @xmath60 , but can range anywhere from a @xmath61 to as high as @xmath62 depending on the level of ionization and the temperature . in ionization equilibrium , a carbon or nitrogen ion that absorbs the ly@xmath0 line
will be photoionized , and will eventually recombine to form emission lines and continua .
in a recombination - dominated plasma at a temperature of @xmath63 , for example , calculations show that approximately 50% of the total number of recombinations produce one of the lyman lines . under most circumstances ,
the compton cross section is roughly two orders of magnitude below the continuum absorption cross section .
this implies that lines photons are preferentially absorbed before they are compton scattered and , therefore , continuum absorption is the dominant mechanism that competes with the line processes .
compton scattering effects will , however , be important in extremely highly - ionized regions , where only trace abundances of bound ions can survive or in low - metalicity plasmas with abundances @xmath64% solar .
these will be ignored entirely in the present work .
to illustrate the importance of line overlap , we assume an infinite , symmetric slab with a pre - specified source distribution of line photons . in reality , dynamical effects , such as velocity gradients and turbulence , as well as inhomogenieties that exist , for example , in accretion disks complicate the problem , but we ignore them in the present study .
although resonance line scattering , in general , produces polarized light , we assume that the radiation field is unpolarized ( cf . , @xcite ) . to simpify the computational procedure even further , the medium is assumed to be isothermal with a constant level of ionization
. we solve the radiative transfer equation in the frequency range near the ly@xmath0 lines by properly accounting for scattering and absorption by the various overlapping lines , as well as the presence of a background continuum opacity source . using the emergent intensity
, we then compute the escape probabilities of the original ly@xmath0 line photons and conversion efficiencies into the lines and photoelectric absorption as functions of the total optical depth of the slab .
the radiative transfer equations for rays propagating in the @xmath65 and @xmath66 directions ( @xmath67 , where @xmath68 is the angle from the normal @xmath69-axis ) in a plane - parallel medium can be written as , @xmath70\ ] ] where @xmath68 is the angle measured from the @xmath69-direction , which we define to be normal to the slab .
the right - hand - side is summed is over all possible sources of opacity and emissivity in the slab .
we express the frequency in terms of the frequency shift from the reference line center at frequency @xmath71 in units of the doppler width @xmath72 , i.e. , @xmath73 , and write the line absorption coefficient ( @xmath74 ) as @xmath75 .
since most x - ray resonance lines have large radiative decay rates ( @xmath76 ) , the population of the upper level is usually negligible compared to that of the ground state . denoting the transition oscillator strength as @xmath77 and the lower - level density as @xmath78 ( @xmath79 ) , we can then write , @xmath80 and @xmath81 is the normalized voigt profile , @xmath82 adopting the notation of @xcite , we define @xmath83 to be the mean optical scale , which increases in the negative @xmath69-direction through the slab .
therefore , @xmath84 , which is related to the line center optical depth @xmath85 by , @xmath86 .
we assume that the continuum is purely absorbing with zero local emissivity .
the source function then contains only the line terms , and can be written as @xcite , @xmath87 where the first term represents the scattering term and @xmath88 is the intrinsic source term .
the quantity @xmath89 is the destruction probability of the line per scattering ( i.e. , @xmath90 for the ly@xmath3 line ) and @xmath91 is the angle - averaged intensity , @xmath92 the scattering term is computed assuming partial redistribution in a voigt profile , in which photons scatter coherently in the atoms frame .
compared to the assumption of complete redistribution , this provides a more accurate representation of line scattering in an optically thick medium where scattering in the damping wings is important .
the frequency redistribution function @xmath93 is defined as the probability that a photon of frequency @xmath94 absorbed by an ion is re - emitted at frequency @xmath95 , and can be written as , @xmath96 - \tan^{-1}\left[\frac{{\rm{min}}(x , x ' ) - x}{a}\right ] \right\ } du,\ ] ] which was first derived by @xcite ( see , also @xcite , and references therein ) .
accounting for the angular redistribution makes negligible difference @xcite and is , therefore , ignored .
the intrinsic source term is frequency independent and can be produced , for example , through collisional excitation and or radiative cascades following recombination .
the exact nature of the excitation mechanism is not important for our present purposes . the transfer equation for multiple overlapping lines
can then be written as , @xmath97,\ ] ] where @xmath98 represent the absorption coefficients normalized to that of the reference line at @xmath99 , and are assumed to be independent of depth .
therefore , the first , second , and third terms on the right hand side of eq .
[ eq : rte2 ] represent absorption , scattering , and intrinsic emission , respectively .
the transfer equation is solved using the feautrier method @xcite , which requires a discretized version of eq .
[ eq : rte2 ] in angular , frequency , and optical depth variables . defining the total opacity @xmath100 ,
where @xmath5 is the continuum opacity relative to the reference line opacity ( @xmath101 ) assumed to be independent of frequency across the line , we define a mean - intensity - like quantity , @xmath102,\ ] ] and rewrite eq . [ eq : rte2 ] as a second order differential equation , @xmath103 the presence of temperature and ion abundance gradients will introduce an additional term on the right hand side , which is ignored for simplicity .
we assume a symmetric slab and specify boundary conditions on the surface and at the midplane of the slab . assuming no incident radiation on either boundaries , they can be written as , @xmath104 on the surface and , @xmath105 at the midplane .
these boundary conditions are also discretized using a second order method described by @xcite .
the frequency integral is represented as a quadrature sum using simpson s formula with @xmath106 points depending on the temperature and the total optical depth of the slab .
we use a 2-point gaussian quadrature for the angular integral in the interval @xmath107 ( @xmath108 ) .
the optical depth is discretized into 12 points per dex .
extensive testing shows that the above discretization scheme yields reliable results .
energy conservation ( see , e.g. , @xcite ) is satisfied to within @xmath109% for most cases studied in this paper , and is no larger than @xmath110% even in the worst cases .
we solve the transfer equation for a source function distributed uniformly within a slab of ionized material for a range of total slab thicknesses .
we define the ly@xmath36 line to be our reference line and measure frequency shifts in units of its doppler width .
the emissitivities of the ly@xmath36 and ly@xmath37 lines are assumed to be proportional to their respective statistical weights of the upper levels , i.e. , @xmath111 , which is approximately correct for both collision - dominated and recombination - dominated line emission .
the absolute numerical values for the source functions are not important for the present purposes , and without any loss of generality , we can write , @xmath112 the solution to eq . [ eq : rte3 ] [ eq : bc_mid ] yields the frequency - dependent intensity at each optical depth and angular points , which can then be used to compute the mean number of scatterings for each of the lines , their mean optical scales , and the escape probabilities .
given the normalization condition ( eq . [ eq : src_norm ] ) , the fraction of the original photons that are absorbed by a transition with destruction probability @xmath113 can be expressed in terms of the mean number of scatterings @xcite . the fraction absorbed and destroyed by line @xmath114 can be written as , @xmath115 where @xmath116 is the mean number of scatterings .
similarly , the fraction of photons absorbed in the continuum can be expressed in terms of the mean optical scale @xmath117 @xcite , i.e. , @xmath118 finally , the fraction of photons that escape through both faces of the slab is simply , @xmath119 where @xmath120 is the emergent intensity on the surface .
fractional conversion efficiencies as a function of the total threshold optical depth at the edge ( @xmath121 ) at temperatures of @xmath46 and @xmath122 are shown in figures [ fig : yield10 ] and [ fig : yield50 ] , respectively .
the underlying continuum opacity is assumed to be @xmath123 . at @xmath46 ,
the line - center cross section of the ly@xmath36 line is @xmath124 and is a factor of @xmath125 times larger than the corresponding continuum absorption cross section at threshold , which is assumed to be @xmath126 for @xcite . for @xmath127 and @xmath46 ( figure [ fig : yield10 ] ) ,
only @xmath128% of the ly@xmath0 photons escape the medium , while @xmath128% are absorbed by and @xmath129% are absorbed by the background continuum .
the total line - center opacity in the ly@xmath36 line is @xmath130 .
we find the mean number of scatterings for the ly@xmath36 line to be @xmath131 , which implies that thermalization through particle collisions is relevant if the gas density is higher than @xmath132 ( see 2 ) . at a temperature of @xmath133 ,
the overlap between the and lines is larger , and the conversion to is enhanced relative to the amount absorbed in the continuum .
the temperature dependence of the conversion efficiencies can be understood from the curves plotted in figure [ fig : yield_kt ] . at low temperatures ,
the ly@xmath3 line lies in the wings of the ly@xmath0 lines and so the conversion process is relatively inefficient . as the temperature
is increased , the amount of overlap between the lines increases rapidly up to a temperature of @xmath134 , which is where the separation between the ly@xmath3 and ly@xmath36 lines is approximately equal to one doppler width .
this temperature is representative of a photoionized plasma at this level of ionization . above this temperature ,
the amount of line overlap increases more slowly than the temperature dependence of the line cross section , which decreases approximately as @xmath135 , and so the efficiency drops again with temperature .
finally , we show in figure [ fig : yield2 ] the effect of increasing the overlapping continuum opacity on the conversion efficiencies . from the derived conversion efficiencies , one can estimate the emergent intensities of each of the lines for a given total optical depth . at @xmath136 and a total optical depth of @xmath127 , approximately 40% of the original ly@xmath0
photons escape through the surfaces of the slab , 40% of then are absorbed by , and the remaining 20% are absorbed in the continuum ( see figure [ fig : yield10 ] ) . assuming that the medium is very highly ionized consisting only of h - like carbon , nitrogen , and oxygen , and further assuming that the emission coefficient ratios of the ly@xmath0 lines are proportional to their abundance ratios , ( i.e. , cno ly@xmath0 relative line intensities of @xmath137 )
, the emergent radiation field will have an ly@xmath0 line intensity of @xmath138 , a ly@xmath0 intensity of @xmath139 , and a carbon line intensity of @xmath140 . here
, we have simply assumed that 33% of the photons absorbed by result in the ly@xmath0 line and 50% of the photons absorbed by produce the ly@xmath0 line ( see 2 ) .
therefore , the oxygen to nitrogen line ratio decreased from 7.7 to 1.5 ( a factor of @xmath110 ) and the oxygen to carbon line ratio decreased from 2.2 to 0.73 ( a factor of @xmath16 ) simply due to finite opacity effects .
we do not attempt to compute the global emergent spectrum , which requires a self - consistent calculation of the ly@xmath0 line source functions including a complete set of ions and a realistic ionization balance calculation . to accurately compute the emergent cno line intensities
, one must properly account for all of the processes that increase the oxygen line source function .
the forest of iron l - shell line emission shortward of the edge at @xmath141 is particularly important and is likely to be the dominant contributor to the source function .
a more detailed investigation will be presented in a future article .
we have identified a pair of overlapping resonance lines , which alters the intensities of some of the most prominent lines in x - ray spectra of cosmic sources under optically thick conditions , analogous to the
bowen fluorescence mechanism . a non - negligible fraction of the ly@xmath0 photons can be converted into emission lines , which can be misinterpreted as an anomalous cno abundance pattern . in general , oxygen emission lines
will be suppressed , while both nitrogen and carbon lines will be enhanced compared to those emitted under optically thin conditions .
as briefly mentioned earlier , detections of non - solar cno abundance ratios have been made in a variety of sources using soft x - ray data obtained with the reflection grating spectrometer onboard _ xmm - newton _ and the _ chandra _ low energy transmission grating spectrometer ( see , e.g. , @xcite ) . in all cases ,
the nitrogen lines are observed to be stronger than expected relative to the oxygen lines .
the carbon lines , on the other hand , are stronger in some sources and weaker in others .
although it is highly likely that non - solar overabundances play a major role in producing the anamolous line fluxes , it is also possible that a substantial part of it is due to the bowen mechanism described in this paper .
detailed modeling and careful inspections of other lines in the observed spectrum are required to break the degeneracy between overabundance and radiative transfer effects .
dynamical effects , such as turbulent motion and velocity gradients , which are not included in the present calculations , may reduce or enhance the conversion efficiency of ly@xmath0 into depending on the geometrical configuration of the ionized medium .
turbulence , for example , can suppress the efficiency if the turbulent length scale is much smaller than the mean free path of the line photons . in the presence of velocity gradients ,
the efficiency can be enhanced in the direction where the ly@xmath3 line is shifted into the ly@xmath0 line core .
these effects coupled with density inhomogeneities will complicate the problem further and must be studied carefully on a case - by - case basis .
there are a few other potentially important line overlaps in the x - ray band that are worth mentioning .
for example , one of the three brightest lines at wavelengths of 17.096 , 17.051 , and 16.780 typically referred to as the m2 , 3 g , 3f lines @xcite , respectively , coincide in wavelength with the high-@xmath142 series lines of at 17.10068 ( @xmath143 ) , 17.02584 ( @xmath144 ) , etc . , and 16.786 ( @xmath145 ) . in this case , dominates the line opacity , while contributes most of the intrinsic source function .
another example is the overlap between the lines at 16.071 and 16.004 @xcite with the ly@xmath5 lines at 16.00666 and 16.00552 .
these examples are certainly not exhaustive and it is highly likely that they must all be accounted for collectively for accurately predicting the global x - ray spectrum of optically thick sources .
the author thanks roger blandford , steven kahn , and duane liedahl for valuable discussions , and mingfeng gu for his help with the flexible atomic code .
this work was supported by nasa through _ chandra _ postdoctoral fellowship award number pf1 - 20016 issued by the _ chandra _
x - ray observatory center , which is operated by the smithsonian astrophysical observatory for and behalf of nasa under contract nas8 - 39073 .
anders , e. & grevesse , n. 1989 , , 53 , 197 auer , l. 1967 , , 150 , l53 bautista , m. a. & kallman , t. r. 2001 , , 134 , 139 bowen , i. s. 1934 , , 46 , 146 bowen , i. s. 1935 , , 81 , 1 branduardi - raymont , g. , sako , m. , kahn , s. m. , brinkman , a. c. , kaastra , j. s. , & page , m. j. 2001 , , 365 , l140 brinkman , a. c. , kaastra , j. s. , van der meer , r. l. j. , kinkhabwala , a. , behar , e. , kahn , s. m. , paerels , f. b. s. , & sako , m. 2002 , , 396 , 761 brown , g. v. , beiersdorfer , p. , liedahl , d. a. , widmann , k. , & kahn , s. m. 1998 , , 502 , 1015 brown , g. v. , beiersdorfer , p. , liedahl , d. a. , widmann , k. , kahn , s. m. , & clothiaux , e. j. 2002 , , 140 , 589 drake , g. w. f. 1986 , , 34 , 2871 drake , s. a. & ulrich , r. k. 1980 , , 42 , 351 eastman , r. g. & macalpine , g. m. 1985 , , 299 , 785 feautrier , p. 1964 , c. r. acad .
paris , 258 , 3189 ferland , g. j. , korista , k. t. , verner , d. a. , ferguson , j. w. , kingdon , j. b. , & verner , e. m. 1998 , , 110 , 761 garcia , j. d. & mack , j. e. 1965 , journal of the optical society of america , 55 , 654 gu , m. f. 2003 , , 582 , 1241 harrington , j. p. 1972
, , 176 , 127 henyey , l. g. 1941 , proc .
, 26 , 50 hummer , d. g. 1962 , , 125 , 21 hummer , d. g. 1969 , , 145 , 95 hummer , d. g. & kunasz , p. b. 1980 , , 236 , 609 ivanov , v. v. 1973 , transfer of radiation in spectral lines ( nbs special publication , washington : us department of commerce , national bureau of standards , 1973 , english language edition ) jimenez - garate , m. a. , hailey , c. j. , herder , j. w. d. , zane , s. , & ramsay , g. 2002 , , 578 , 391 johnson , w. r. & soff , g. 1985 , atomic data and nuclear data tables , vol .
33 , p.405 , 33 , 405 kallman , t. & mccray , r. 1980 , , 242 , 615 kallman , t. r. & mccray , r. 1982 , , 50 , 263 kinkhabwala , a. , sako , m. , behar , e. , kahn , s. m. , paerels , f. , brinkman , a. c. , kaastra , j. s. , gu , m. f. , et al .
2002 , , 575 , 732 lee , h .- w . , blandford , r. d. , & western , l. 1994 , , 267 , 303 leutenegger , m. a. , kahn , s. m. , & ramsay , g. 2002 , , 585 , 1015 mason , k. o. , branduardi - raymont , g. , ogle , p. m. , page , m. j. , puchnarewicz , e. m. , behar , e. , c ' ordova , f. a. , davis , s. , et al .
2003 , , 582 , 95 milkey , r. w. , shine , r. a. , & mihalas , d. 1975 , , 202 , 250 netzer , h. , elitzur , m. , & ferland , g. j. 1985 , , 299 , 752 ogle , p. m. , brookings , t. , canizares , c. r. , lee , j. c. , & marshall , h. l. 2003 , , 2003 , 402 , 849 seaton , m. j. 1959 , , 119 , 90 shine , r. a. , milkey , r. w. , & mihalas , d. 1975 , , 201 , 222 unno , w. 1955 , , 7 , 81 verner , d. a. , ferland , g. j. , korista , k. t. , & yakovlev , d. g. 1996 , , 465 , 487 weymann , r. j. & williams , r. e. 1969 , , 157 , 1201 | radiative transfer effects due to overlapping x - ray lines in a high - temperature , optically - thick , highly - ionized medium are investigated .
one particular example , where the ly@xmath0 doublet ( @xmath1 @xmath2 ) coincide in frequency with the ly@xmath3 lines ( @xmath4 @xmath2 ) is studied in detail to illustrate the effects on the properties of the emergent line spectrum .
we solve the radiative transfer equation to study the energy transport of resonance line radiation in a static , infinite , plane - parallel geometry , which is used to compute the destruction / escape probabilities for each of the lines for various total optical thicknesses of the medium , as well as destruction probabilities by sources of underlying photoelectric opacity .
it is found that a large fraction of the ly@xmath0 line radiation can be destroyed by , which can result in an reversal of the ly@xmath0/ ly@xmath0 line intensity ratio similar to what may be seen under non - solar abundances .
photoelectric absorption by ionized carbon and nitrogen can also subsequently increase the emission line intensities of these ions .
we show that line ratios , which are directly proportional to the abundance ratios in optically thin plasmas , are not good indicators of the true cno abundances .
conversely , global spectral modeling that assumes optically thin conditions may yield incorrect abundance estimates when compared to observations , especially if the optical depth is large .
other potentially important overlapping lines and continua in the x - ray band are also identified and their possible relevance to recent high resolution spectroscopic observations with _
chandra _ and _ xmm - newton _ are briefly discussed . |
from the times of the celebrated greenwood and williamson ( 1966 ) paper , we know that surfaces that have been worn or abraded show a non - unique gaussian distribution .
indeed , fig.6 of gw ( adapted here as fig.1 ) shows in a gaussian paper the distribution of summit heights in a surface of mild steel which had been abraded and then slid against copper , which resembles a `` bimodal gaussian '' .
more precisely , greenwood & wu 2001 returned to that very surface and commented that fig.6 of gw paper _ `` shows an abraded surface which is largely flat but where the heights of the upper 80% may be regarded as gaussian .
( we may note that ` proper ' statistical tests would not reveal this fact)''_. @xmath0{cc}{\includegraphics [ natheight=7.135600 in , natwidth=5.874700 in , height=3.5968 in , width=2.9654 in ] { g66.jpg } } & \end{array } $ ] fig.1 - height distribution of an abraded surface ( adapted from greenwood and williamson 1966 ) . on more careful consideration
therefore , the surface had been largely worn out , but this may not be always the case . today , we can gather a lot more insights into the process of wear or abrasion from the models developed , using gw description of roughness , in the context of chemical mechanical polishing used for planarization of integrated circuits ( borucki , 2002 , borucki _ et al .
_ 2004 , shi & ring 2010 ) . indeed , polishing causes high asperities wearing faster than low asperities .
assume an original distribution of asperities , @xmath1 which is taken to be gaussian ( but we postpone the term @xmath2 into the integrals for easy of notation)@xmath3 \quad,\quad\overline{\phi}_{0}=\exp\left [ -\frac{z^{2}}{2}\right]\ ] ] where @xmath4 is the rms amplitude of the summit heights .
using archard law for wear and the gw model , an equation similar to a differential hamilton - jacobi equation was obtained for the distribution of asperity heights which tends to develop very high peaks in the tail and indeed in the case when the separation between the polishing pad and the surface is kept constant at @xmath5 , an analytical solution is developed with the method of characteristics as ( borucki et al 2004 , shi & ring 2010 ) @xmath6 \quad,\quad z > c\\ & = \frac{w}{2\sqrt{z - c}}\left ( t+\frac{2}{w}\sqrt{z - c}\right ) \phi _ { 0}\left [ z+\frac{w^{2}}{4}t^{2}+\frac{w}{2}t\sqrt{z - c}\right ] \quad,\quad z > c\end{aligned}\ ] ] where @xmath7 , with @xmath8 and @xmath9 is the wear constant , while @xmath10 is sliding speed .
further , @xmath11 is elastic modulus of materials , and @xmath12 the radius of asperity summits .
as borucki et al ( 2004 ) clearly state:- `` this solution develops an integrable singularity at @xmath13 representing the portion of the pad - surface that has been _ worn smooth _ by the wafer .
after a sufficiently long but finite time , this singularity converges to a @xmath14distribution with its amplitude approaching a limiting value given by the fraction of the pad - surface originally protruding above the wafer height '' .
we need to take this carefully in consideration when developing a contact mechanics model . before we do that ,
let us move to dimensionless form , @xmath15 \quad,\quad\overline{z}>\overline{c}\ ] ] where we have introduced a dimensionless time@xmath16 @xmath0{cc}{\includegraphics [ height=2.2857 in , width=3.659 in ] { phi.eps } } & \end{array } $ ] fig.2 .
the distribution of asperity heights due to wear in borucki s solution @xmath17 .
@xmath18 , and @xmath19
. the delta function can not be represented for obvious reasons .
this function is plotted in fig.2 for representative dimensionless time @xmath19 , and compared to the original gaussian tail .
we can integrate analytically the difference between the borucki distribution , and the original gaussian one , to find the fraction of asperities worn out @xmath20\ ] ] which we shall assume , in the absence of a better hypothesis , all lye on the @xmath21 , and , perhaps an even stronger assumption , maintain the same radius .
obviously if special experimental setups will permit to have reasonably simple alternative assumptions , they could be readily incorporated . therefore , the so - modified borucki distribution is @xmath22 + f\left ( \overline { c},\overline{t}\right ) \delta\left ( \overline{z}-\overline{c}\right ) \quad,\quad\overline{z}>\overline{c}\ ] ] where @xmath23 is the classical delta function .
by developing the usual gw treatment for number of asperities in contact , area and load , integrating for the distribution of asperity heights , we obtain for the compression @xmath24@xmath25 we can define the gaussian case integrals as @xmath26 whereas the borucki version `` before the modification '' @xmath27 and after the modification , in considering the @xmath28 functions , additional contributions lead to heaviside functions , @xmath29 further , writing @xmath30 and @xmath31 we can rewrite our final results as@xmath32
the results are quite as expected .
fig.3 plots the integrals @xmath33 for the same case of fig.2 . starting with the integral giving the number of asperities in contact ,
it is clear that , at the `` jump '' @xmath18 , we recover the number of asperities in the unworn profile , by construction . upon increasing the wear time ,
the tail is worn out increasingly .
looking now at fig.3b , this integral is proportional to the contact area , and , as in fig.3c , a change of slope results in moving across the jump , but not an actual jump as in the number of asperities in fig.3a .
@xmath0{cc}{\includegraphics [ height=2.2857 in , width=3.659 in ] { i0.eps } } & ( a)\\{\includegraphics [ height=2.2857 in , width=3.659 in ] { i1.eps } } & ( b)\\{\includegraphics [ height=2.2857 in , width=3.659 in ] { i3-2.eps } } & ( c ) \end{array } $ ] fig.3 .
the integrals @xmath33 ( a , b , c , respectively ) .
@xmath18 , and @xmath19 .
we can define a mean pressure on the contact@xmath34 which for the unworn given gaussian system , would be relatively constant with separation .
fig.4 shows that the pressure is indeed more or less constant for any system , despite the constant changes with time , decreasing as long as the process continues this is mainly due to the fact that the `` effective '' roughness is reduced in the tail of the distribution , because of wear . on the other hand ,
if the contact moves to separations lower than @xmath18 , then the mean pressure tends to restore rather quickly its original gaussian value it does nt do immediately so because the asperities have now moved to a lower height , and therefore , with respect to the unworn case , they are compressed of a much lesser amount , resulting in less pressure .
obviously this suggests that for the procedure to really work at constant separation @xmath35 , the total load would need to decrease according to the model prediction .
@xmath0{cc}{\includegraphics [ height=2.2857 in , width=3.659 in ] { i3-2overi1.eps } } & \end{array } $ ] fig.4 . the ratio of the integrals @xmath36 which indicates mean pressure in the contact areas .
@xmath18 , and @xmath19 .
we have shown in a simple case , that wear or abrasion leads to continuous change of roughness parameters , which affect the response of the rough contact . in cases when the process is selectively acting on part of the asperity distribution ,
the effects are clearer , and a bimodal distribution results .
it may be therefore be an oversimplification to consider surfaces gaussian , as if coming simply from a single distribution .
borucki , l. mathematical modeling of polish rate decay in chemical - mechanical polishing .
( 2002 ) 105114 | since the time of the original greenwood & williamson paper , it was noticed that abrasion and wear lead to possibly bimodal distribution of asperity height distribution , with the upper tail of asperities following from the characteristics of the process . using a limit case solution due to borucki for the wear of an originally gaussian distribution , it is shown here that the tail is indeed always gaussian , but with different equivalent parameters . therefore , if the wear process is light , one obtains a bimodal distribution and both may affect the resulting contact mechanics behaviour . in this short note
, we illustrate just the main features of the problem .
we conclude that it is an oversimplification to consider surfaces gaussian . |
stellar bowshock nebulae are arcuate structures created by the interaction between stellar winds and the surrounding interstellar medium ( ism ) where the relative velocity between the two is supersonic @xcite .
a shock front forms at the interface of the high speed stellar wind and the ambient interstellar medium . the swept up interstellar dust and gas ahead of the high - velocity star forms an arc - like feature preceding the peculiar motion of the star .
this material is most clearly visible in the mid - infrared and sometimes in optical emission lines such as h@xmath2 and @xcite .
most bowshock nebulae are observed around high - velocity massive stars , but they have also been identified preceding high - velocity pulsars @xcite , red supergiants @xcite , and associated with proplyds in the orion nebula @xcite .
@xcite derived an exact analytic solution for this interaction by balancing the ism pressure and ram pressure of the stellar winds .
based on this model , @xcite used computer simulations to demonstrate that stable bowshocks can be formed for a variety of interaction scenarios over a range of relative velocities , ism densities , and stellar wind momentum fluxes . a significant proportion ( 10@xmath325% ) of early - type stars
reside outside of stellar clusters @xcite . because of their short main - sequence lifetimes these stars , known as runaway stars , must be moving at high peculiar velocities , typically > 30 km s@xmath4 @xcite .
@xcite found that o and b type stars ( ob stars ) comprise 50% of all runaway stars , and these ob runaways are dominated by o type stars ( 12:1 ) .
two general scenarios that generate runaway stars have been proposed : close dynamical interactions of either single stars @xcite or binary - binary interactions @xcite , and ejection from close binary systems when the more evolved member undergoes a core - collapse supernova ( ccsn ; * ? ? ?
* ; * ? ? ?
@xcite analytically simulated the dynamical interactions of small clusters of 56 50 m@xmath5 stars and found that 217% of these stars were ejected from the cluster with velocities in excess of 35 km s@xmath4 .
@xcite concluded that dynamical few - body ejections of ob stars likely occur in the first @xmath61 myr of the cluster s history before the radiation of the ob stars expels the mass - dominant gas component of the cluster , causing the cluster to expand .
@xcite simulated binary - binary gravitational interactions and found a maximum ejection velocity of 700 km s@xmath4 for 60 m@xmath5 stars and 1400 km s@xmath4 for 14 m@xmath5 stars . though the initial multiplicity fraction of ob stars is unknown ,
studies have shown that the fraction is likely higher than 60% in stellar clusters , meaning that at least 7590% of massive stars are in a multiple system and have the potential to participate in few - body interactions @xcite .
@xcite first proposed that stars that form in close binary systems could be ejected when one of the components explodes as a supernova .
@xcite hydrodynamically simulated the effects of asymmetric ccsn in close binary systems and determined that the less - evolved companion could survive with minimal disruption and be ejected at runaway velocities .
@xcite used monte carlo simulations to determine the necessary impulse to account for the observed velocity distribution of pulsars and found by extension that the necessary forces could accelerate the surviving ob companions to velocities up to 100 km s@xmath4 .
@xcite expanded upon this work with new simulations and found that early - type b stars ( 10 m@xmath5 ) could be accelerated up to @xmath1320 km s@xmath4 under ideal conditions with increasing maximum velocities for lower mass stars ( up to @xmath11050 km s@xmath4 for a 0.90 m@xmath5star ) .
it is likely that both few - body interactions and supernovae in binary systems combine to generate the observed population of runaway stars . using hipparcos proper motions @xcite traced the motions of two known runaway stars , ae aur and @xmath0 col ( o9.5v and o9.5b / b0 , moving in opposite directions at 100 km s@xmath4 ) to a common origin @xmath12.5 million years ago near the location of the binary pair @xmath7 ori .
they proposed that the two runaways may have been ejected in a binary - binary interaction .
@xcite identified two o - type runaways that may have been ejected from the star cluster ngc 3603 in a single star - binary interaction when the single star captured one binary member and ejected the other with both systems being accelerated to high peculiar velocities relative to the cluster .
the new binary system later merged into the single blue straggler star observed .
@xcite used hipparcos proper motions to trace @xmath8 oph and the pulsar j1932 + 1059 back to the same region of the upper scorpius star forming region @xmath11 myr ago , which could indicate that the two were once members of a common binary prior to the supernova creating the pulsar .
@xcite identified eight additional binary - binary and 16 binary - supernova ejection candidates by extrapolating their proper motions .
a fraction of the ejected stars in the @xcite simulations accreted material from the ccsn , and these simulations predict that atmospheric chemical enrichments may be observable in the runaway stars if mixing is inefficient .
several runaway stars @xcite have been observed with significant @xmath2-element enrichment , consistent with this prediction .
hybrids of these ejection scenarios may also be possible .
@xcite suggested that a two - step scenario involving a dynamically ejected binary pair could accelerate a single runaway star following a ccsn . because of the two kicks accelerating the star in this scenario , the star s observed proper motion vector may not point back to a cluster .
@xcite examined a sample of known runaway stars and found that 4@xmath32% do not have proper motions that can be extrapolated back to a known cluster or association .
hypervelocity stars ( hvss ) are a special class of runaway stars with extremely high space velocities , typically defined as being greater than 400 km s@xmath4 @xcite .
@xcite first predicted hypervelocity stars when modelling the close interaction of binary star systems with a supermassive black hole ( smbh ) , which produced stars with velocities in excess of 1000 km s@xmath4 .
@xcite proposed two additional mechanisms : gravitational interaction between a pair of single stars with a extreme mass ratio or stellar acceleration by a binary smbh system .
however , they predicted that the hvs generation rate by single star interactions is likely low enough to be undetectable due to the extremely small impact parameter required .
@xcite reported the discovery of the first hvs in the galactic halo , sdss j090745.0 + 024507 , a main sequence b star @xcite which remains the fastest observed hvs with a heliocentric radial velocity of [email protected] km s@xmath4 @xcite .
@xcite assembled the most complete catalog of hvss with 24 confirmed objects and several additional candidates .
several simulations @xcite have demonstrated that it is unlikely that hvss can support visible bowshock nebulae due to their high velocities inhibiting the accumulation of material in the leading shock .
two general classes of bowshock nebulae are generally recognized : those supported by runaway stars , and `` in - situ '' bowshocks supported by a star overrun by an outflow of hot gas from a star - forming or region .
@xcite used optical emission line imaging to catalog the first bowshock nebulae which appeared as `` distorted interstellar bubbles '' .
these nebulae were observed around the prototypical runaway @xmath8 oph and around the star ll ori situated in an outflow from @xmath9 ori in the orion nebula .
@xcite cataloged six arcuate nebulae around the star - forming regions m 17 and rcw 49 and identified these in - situ shocks as a distinct class .
the associated physics of these in - situ shocks are similar to runaway bowshocks when considered within the rest frame of the stellar source . using 60 @xmath0 m images from the infrared astronomical satellite ( @xmath10 ) @xcite @xcite compiled the first catalog of bowshock nebulae .
a more complete sample gathered from @xmath10 images found a total of 58 bowshock nebula candidates around 188 runaway ob stars @xcite .
subsequent re - analysis of @xmath10 all - sky images concluded only 19 of the 58 candidate were bowshock nebulae with 2 additional questionable candidates @xcite .
the increased angular resolution of infrared surveys conducted by the spitzer space telescope ( @xmath11 ) and wide - field infrared survey explorer ( @xmath12 ) @xcite , enabled the identification of small collections of bowshock nebulae in both the lmc @xcite and smc @xcite , generated by stars ejected from the star - forming regions ngc 6611 @xcite , cygnus ob2 @xcite , ngc 6357 @xcite , ngc 3603 @xcite , and carina nebula @xcite , and powered by high - mass x - ray binaries @xcite , pulsars @xcite , red @xcite and blue @xcite supergiants , and the a - type star @xmath13 vel @xcite .
@xcite conducted the most extensive search to date for bowshock nebulae and runaway stars by examining @xmath12 archival images in the vicinity of several hundred runaway star candidates identified by @xcite .
the most recent release of their e - boss ( extensive bow shock survey ) catalog @xcite contains a compilation of 73 bowshock nebulae candidates found in their visual search and those previously identified in the literature .
based on their work , it appears that 510% of runaway stars support bowshock nebulae , though some earlier studies have suggested higher rates of @xmath115% @xcite or as high as @xmath130% @xcite .
the present paper is the first in a series presenting and analyzing a new catalog of candidate bowshock nebulae .
our group has taken the opposite approach of the e - boss survey @xcite by conducting a comprehensive visual inspection of existing space - based infrared surveys to identify candidate bowshock nebulae and runaway stars along the galactic mid - plane . in this paper , we report the identification of 708 candidate bowshock nebulae , 659 of which have not been previously identified in the literature .
this catalog constitutes the largest collection of bowshock nebula candidates by an order of magnitude . in @xcite , we present the results of spectroscopic follow - up of nearly 100 targets , that @xmath1495% of these infrared - selected stellar sources are , in fact , early - type massive stars , and that infrared nebular morphology alone is enough to select massive stars with confidence . in section 2 ,
we discuss the selection process and space - based infrared datasets used for identifying candidate nebulae .
section 3 presents the archival infrared images of our candidates and discusses the demographics of our sample including the sky distribution , proper motions , and their surrounding environments .
we summarize our results in section 4 .
our team conducted a visual examination of mid - infrared images from _ spitzer space telescope _
( @xmath11 ; * ? ? ? * ) and the wide - field infrared survey explorer ( wise ; * ? ? ? * ) to locate bowshock nebula candidates .
the @xmath15 data included several wide - area surveys conducted using the infrared array camera ( irac ; * ? ? ? * ) in its 3.6 , 4.5 , 5.8 , and 8.0 @xmath0 m bandpasses along with 24 @xmath0 m data from the multiband imaging photometer for spitzer ( mips ; * ? ? ? * ) .
the @xmath15 beam size at these bands is 1.66 , 1.72 , 1.88 , 1.98 and 6 fwhm , respectively .
survey data included images obtained as part of the galactic legacy midplane survey extraordinaire i and ii ( glimpse , glimpse ii ; * ? ? ? * ) and mipsgal @xcite programs , the the @xmath16 legacy survey of the cygnus - x complex @xcite , and the @xmath16 mapping of the outer galaxy ( smog ; * ? ? ? * ) .
together these @xmath11 surveys cover slightly more than 304 square degrees of the galactic plane a large continuous strip at @xmath17 , @xmath18 ( glimpse and mipsgal ) , a 24 square degree regions near @xmath19 ( cyg - x ) , and a 2@xmath20 10 strip centered on @xmath21=105(smog ) .
our search used the large mosaics at each waveband constructed by the glimpse team that are available at the spitzer science center .
the @xmath12 data include images at the 3.4 , 4.6 , 12 , and 22 @xmath0 m bandpasses which have beamsizes of 6.1 , 6.4 , 6.5 , and 12.0 fwhm , respectively .
our search of the @xmath12 all - sky survey images ( about 460 square degrees of the outer galaxy ) used 1 mosaics in regions along the galactic plane not covered by spitzer and the centre de donnes astronomiques de strasbourg s aladin interface which allows a viewer to rapidly browse large regions of the sky .
given the concentration of bowshock nebulae very near the galactic plane ( section [ loc.sec ] ) , our search of the @xmath12 data concentrated on regions @xmath22 @xmath23 2 , but we also performed a cursory search using the aladin @xcite lite interface up to @xmath22 @xmath23 10 .
only a few objects were discovered beyond @xmath24 @xmath31 .
we searched the @xmath11 mosaics for 1 ) arc - shaped or lima - bean - shaped nebulae at 24 @xmath0 m and/or 8.0 @xmath0 m that had 2 ) a high degree of symmetry with 3 ) at least one prominent star located near the axis of symmetry visible at 4.5 @xmath0 m .
for the majority of candidates , there is a single bright star located along the axis of symmetry ; it constitutes the obvious early - type stellar candidate , even though there is no spectral information available for the vast majority of candidates . in some cases there
are multiple candidate stars within the arc , and in a few cases there are no prominent stellar candidates , especially in the inner galaxy where the stellar density is high and even early type stars may appear faint because they are located at large distances .
our initial visual search team interpreted these three criteria liberally .
the galactic plane exhibits very complex structures at 8.0 and 24 @xmath0 m that vary greatly in surface brightness over small angular scales , producing a large number of initial candidates that do not meet one of more of the three search criteria .
the final decision regarding which objects met all three criteria was ultimately made by the senior author ( hak ) .
relaxing these criteria would substantially increase the candidate list by at least a factor of two .
the search process involved examining color representations of the 4.5 , 8.0 , and 24 @xmath0 m images displayed with a square - root color stretch .
owing to the high dynamic range of the mosaics , it was necessary to use a variable maximum and minimum intensity levels to achieve the needed contrast , depending on the surface brightness in each region . often in regions of large dynamic range several different scalings
were used .
this was necessary to find nebulae superimposed on regions with high background levels ( e.g. , within regions ) .
no quantitative thresholds were used in the search process ; indeed , the rms noise levels are not well defined , varying significantly on small spatial scales where regions of very high and very low surface brightness appear adjacent to one another .
the maximum angular size of a detectable candidate is limited by the size of the survey mosaics at about 2 degrees ; in practice , there are no candidates larger than 8.6 .
the angular resolution of the @xmath11/@xmath12 surveys , described above , sets the lower limit on the angular size of detectable candidates .
two or more team members searched each region of the plane . in comparing candidate lists identified by each searcher
we concluded that there was at least a 5075% overlap between the candidate lists .
each searcher identified candidates not found by the others .
we estimate that use of additional searchers could increase the current candidate list , most of this gain coming in areas of the plane dominated by high - surface brightness regions . a search such as this
can never be deemed `` complete '' given the inherent subjectivity of what constitutes `` arc - shaped '' or `` symmetric '' , especially in regions having large changes in surface brightness over small angular scales where genuine bowshock nebulae may be confused with unrelated foreground or background emission .
all @xmath11 objects were discovered initially in the 24 @xmath0 m bandpasses where emission is either thermal in nature or from ( semi-)stochastic heating of small dust grains . only a minority of the bowshock candidates exhibit morphologically similar counterparts in the 8.0 @xmath0 m band , which predominantly includes a broad solid - state emission feature from polycyclic aromatic hydrocarbons ( pahs ) .
however , the 8.0 @xmath0 m band may also include a contribution from very hot dust near the sublimation temperature .
our search of the wise images concentrated on the 22 @xmath0 m ( w4 ) band for nebula identification and the 4.6 @xmath0 m ( w2 ) band for identification of a stellar source .
we find that nearly all candidates identified in the w4 band also had counterparts in the 12 @xmath0 m ( w3 ) band , which may include both pah and hot dust emission .
owing to the superior angular resolution of @xmath11 compared to wise
6.0at 24 @xmath0 m versus 12.0 fwhm at 22 @xmath0 m
we gave preference to the @xmath11 data ( where available ) and did not search the overlapping @xmath12 images . because the @xmath12 beamsize is twice that of the @xmath11 at the longest wavelengths , objects that would appear arcuate in @xmath11 imaging might appear circular in @xmath12 and , therefore , be rejected from the catalog .
this leads to a kind of systematic decrease in the completeness ( at fixed angular sizes of 612 ) for the regions surveyed only with @xmath12 .
similarly , we find a large number of small circular 24 @xmath0 m nebulae near the angular resolution limit of @xmath11 that might appear arcuate if viewed under higher angular resolution .
additionally , bowshock candidates listed in the literature were examined and retained if they met our morphological selection criteria described above .
our survey of the literature included the compilations of ( * ? ? ?
* pe15 ) , ( * ? ? ?
* s15 ) , ( * ? ? ?
* pe12 ) , ( * ? ? ?
* g11 ) , ( * ? ? ?
* k10 ) , ( * ? ? ?
* g08 ) , ( * ? ? ?
* p08 ) , ( * ? ? ?
* cp07 ) , ( * ? ? ?
* n97 ) , and @xcite .
the final collection tabulated here does not include many of the candidates listed in these works , primarily because they fail to meet one of more of the morphological criteria .
table [ bigtable.tab ] lists parameters for the probable central stars of the 709 candidate bowshock nebula .
the first twenty rows appear in the printed journal to give guidance as to the table format and content .
the entire table is available in digital format in the electronic edition .
column 1 is a sequential identification number unique to this work , column 2 is a generic name in galactic coordinates , while columns 3 and 4 list the j2000 right ascension and declination for the putative central star .
column 5 gives the publication where the candidate is first identified as a bowshock candidate .
a `` t '' indicates `` this work '' , and other abbreviations refer to published works , as referenced fully in the table footnotes .
of the 709 tabulated objects , 660 are identified herein as bowshock candidates for the first time .
column 6 lists another common alias for the central stellar source , if one exists in the literature , such as an hd ( henry draper catalog ) designation .
fully 581 of the 709 stellar sources have no alternate designation in the literature , indicating that the vast majority of the objects are faint and highly reddened .
column 7 is a single character designating whether the nebular object has an @xmath11 irac 8.0 counterpart with a similar morphology ( y ) , has no detection above background levels at this waveband ( n ) , or has no data from irac ( -).m band also have counterparts in the wise 12 @xmath0 m band .
therefore , we do not tabulate detections for the wise bandpasses . ]
we find that 130 of the 626 objects having irac data show a nebula at 8.0 @xmath0 m , indicating a probable pah contribution .
column 8 contains a single alpha - numeric code , `` c '' , in cases where the probable central stellar source is uncertain owing to faintness or to the presence of multiple stars of similar magnitude near the nebular axis ; 115 objects fall into this category .
in such cases we adopt the brightest point source nearest the axis of symmetry as the most probable star , but we urge caution with regard to the certainty of this identification .
columns 9 and 10 list the distance @xmath25 ( in arcsec ) from the putative central star to the apsis of the candidate bowshock and the position angle ( in degrees from n toward e ) of this vector in equatorial coordinates .
these values were measured by eye and carry typical uncertainties of 15 and 5 , respectively .
columns 11 and 12 list the 2mass h - band and irac 4.5/wise 4.6 @xmath0 m magnitudes of the central star , obtained from the 2mass @xcite and glimpse / wise catalogs , respectively . in just a handful of cases ( 19 )
, no h - band source is measured , implying an extremely red , highly extinguished object .
column 13 lists an estimate of the k - band extinction to each source , inferred using the [email protected] color and the rayleigh - jean color excess formulation of @xcite .
the final column of table [ bigtable.tab ] provides additional descriptors that characterize the local environment of each stellar candidate .
column 14 contains an alphanumeric code that describes the immediate environment of each candidate as either isolated ( i ; 499instances ) , directly facing a large region within about 10 arcminutes ( fh ; 103 instances ) , facing a bright - rimmed cloud prominent at 8.0 @xmath0 m within several arcminutes ( fb ; 58 instances ) , or situated within a giant region ( h ; 49 instances ) . in some cases two designations could apply , for example , a candidate that lies within an region that also faces a bright - rimmed cloud . in such cases fb
is given preference over h. objects designated fh may be instances of the `` in - situ '' bowshocks @xcite .
these descriptors are qualitative judgments of the candidates immediate environments intended to elucidate possible physical origins for what we regard as potentially different classes of bowshock nebulae .
given the unknown distances of each object and complex sightlines in the galactic plane , physical connections between bowshock candidates and other objects nearby on the sky are necessarily uncertain .
section [ env.sec ] will discuss example objects in each of these five categories in more detail .
figure [ prototypes ] shows a collage of three - color representations for six objects from table [ bigtable.tab ] that may be regarded as prototypical bowshock nebulae .
the angular scale of each panel differs to accommodate the size of each object .
each panel displays a highly resolved , unmistakable arc - shaped nebulae and a bright central star with a substantial proper motion consistent with the nebular bowshock orientation .
these six objects are @xmath8 oph ( g006.2812 + 23.5877 ; upper left ) , ae aur ( [email protected] ; upper right ) , hd136003 ( g322.6802 + 00.9060 ; center left ) , hd150898 ( [email protected] ; center right ) , hd155755 ( g348.7967 + 00.1455 ; lower left ) , hd143275 ( g350.0969 + 22.4904 ; lower right ) , as `` prototypical '' examples of bowshock nebulae drawn from the literature .
we use a color scale that depicts the @xmath11 24 @xmath0 m or @xmath12 22 @xmath0 m image in red , the @xmath11 8.0 or @xmath12 12 @xmath0 m image in green , and the @xmath11 3.6 or @xmath12 3.6 @xmath0 m image in blue .
a white arrow depicts the proper motion vector magnitude and direction .
images of g322.6802 + 00.9060 and g348.7967 + 00.1455 use irac data so the nebula is dominated by 24 @xmath0 m emission ( red ) and appears red , while the other four use @xmath12 data , and the combination of 22 @xmath0 m ( red ) and 12 @xmath0 m ( green ) emission results in a yellowish appearance .
figure [ prototypes ] reveals that only a minority of the objects display the ideal ( `` theoretical '' ) bowshock morphology e.g. , see figure 1 of @xcite or figure 9 of @xcite or figure 3 of @xcite for a variety of viewing inclination angles ) .
the @xmath8 oph nebula ( upper left ) appears flocculent and significantly asymmetric , perhaps because it is highly resolved as the nearest of the bowshocks at a distance of only 140 pc , based on its hipparcos parallax @xcite .
most of the nebulae are , to some degree , asymmetric . in some cases , the star appears slightly displaced from the axis of symmetry .
@xcite discusses such a displacement as the natural result of either a density gradient in the ambient medium transverse to the star s motion or a stellar wind that is anisotropic .
a digital appendix contains images of all 709 candidate objects from table [ bigtable.tab ] in figures a.1 through a.118 using the same color scheme as figure [ prototypes ] .
these figures depict an angular field of view of 2 arcmin square in most cases to facilitate easy angular size comparison , but for some objects a window of 4 or 8 is necessary to include the entire nebula or pertinent features in the immediate environment . for a few objects observed by the @xmath11 irac but not at mips 24 @xmath0 m , we use the @xmath12 22 @xmath0 m image instead . in a just a few instances ,
the images are constructed using the irac 8.0/4.5/3.6 @xmath0 m images . in some panels
a white arrow depicts the proper motion vector magnitude and direction , if proper motion data are available and the measured proper motion is larger than the uncertainties .
the vast majority of the 709 candidate bowshock nebulae have small angular sizes and are , consequently , less resolved than the prototypical examples ; nevertheless , they share the overall shape , colors , and morphological symmetry of the prototypes .
figure [ allcolor180 ] plots the positions of the candidate bowshock nebulae on the galactic plane for the 663 objects that lie in the inner two galactic quadrants at longitudes @xmath27 .
colors denote the extiction values estimated from the [email protected] @xmath0 m color excess : cyan@xmath28@xmath232 ; green@xmath29@xmath28@xmath30 ; red@xmath31@xmath28@xmath32 ; magenta@xmath28@xmath33 ; white no data .
the points appear superimposed on a three - color representation of the galactic plane with the infrared astronomical satellite ( @xmath10 ) maps at 25/60/100 @xmath0 m in blue / green / red , respectively .
the distribution of objects is non - uniform , even allowing for the fact that they were selected from surveys covering only a portion of the plane , principally within one degree of @xmath34=0 and within 60 of galactic center .
for example , there is an overdensity of objects in the vicinity spiral arm tangencies near @xmath21=2530 where the path length through spiral arms is long .
the cygnus - x region near @xmath21=79 also exhibits an abundance of objects , probably the result of having multiple nearby ( 1.21.5 kpc ) regions of massive star formation .
the carina star forming region near @xmath21=287 also exhibits a large number of candidates .
a close examination of figure [ allcolor180 ] further reveals that objects having similar extinctions cluster together on the sky , a probable signature of clusterings of massive stars at similar distances and reddennings .
figure [ histb ] is a histogram of the galactic latitudes for the 561 candidates that lie within the footprint of the glimpse and mipsgal surveys where coverage is complete : @xmath35,@xmath36 .
the histogram is strongly peaked near @xmath34=0 , indicating that very few objects meeting our selection criteria are likely to be found at larger latitudes .
the angular scale height for these objects is approximately 0.6 from mid - plane , identical to the scale height for interstellar bubbles and regions @xcite .
figure [ pamap ] displays the positions of bowshocks candidates at @xmath37 ( _ upper panel _ ) and @xmath38 ( _ lower panel _ ) .
symbol color and the background three - color iras image is the same as in figure [ allcolor180 ] .
vector directions indicate the orientation of the bowshock nebulae as determined by the ir morphology .
this figure further illustrates that objects with similar extinctions often cluster together on the sky .
furthermore , there are also clusters of objects that have similar orientations .
one notable example of this is at the base of the giant gas pillar in m 16 near @xmath21=16.8 , @xmath34=0.6 where four objects all point toward the source of radiation that has sculpted the pillar .
figure [ m16four ] depicts four candidate bowshock nebulae ( g016.8118 + 00.6679 , g016.8760 + 00.6456 , g016.8930 + 00.6800 , and g016.9278 + 00.6323 ) in the same field , all pointing in the same direction as the pillars .
the central stars of the bowshocks are hd 168183 ( o9.5iii+b4v ) , bd@xmath2613 4936 ( b0e ) , bd@xmath2613 4934 ( b1vp ) , and bd@xmath2613 4937 ( b1.5v ) , respectively , all early - type stars .
bd@xmath2613 4937 has a measured proper motion directly _ away _ from the source of radiation in this young star forming region .
these are excellent candidates for `` in - situ '' bowshocks @xcite formed by the interaction between the star s own wind and an outflow from the m 16 region . in regions like this one , bowshocks exhibit a similar orientation not because they share a common space velocity but because a common source of external momentum flux impinges upon them .
there are many other possible `` in - situ '' bowshocks in our sample , and these are flagged by fh in table [ bigtable.tab ] . in order to test the hypothesis that bowshock orientations in this field and others may be correlated on small angular scales
, we computed the difference in position angle between each pair of the n=709 nebulae ( n@xmath20(n@xmath261)/2 unique pairings ) and then plotted the mean @xmath39pa versus angular separation in 8 bins in figure [ structure ] . for comparison to the null hypothesis ( that the position angles are uncorrelated ) we randomly shuffled the position angles among the objects 1000 times and computed the mean @xmath39pa and dispersion in each bin .
figure [ structure ] plots the data ( blue points ) and the results of the random monte carlo iterations , where the heavy black line depicts the mean difference in position angle .
the thin solid and dashed lines depict the 1@xmath40 and 2@xmath40 deviations about the mean . at small angular separations
@xmath2320 the two blue points lie near the 2@xmath40 envelope , while at larger angular separations the points scatter about the mean expected in the case of random orientations .
this demonstrates that bowshock positions angles are correlated on small angular scales , as would be expected if some are produced by large - scale external influences such as outflows from nearby regions .
we also investigated whether the bowshock nebulae display preferred orientations in the galactic coordinate system .
figure [ histpa ] displays a histogram of the morphological position angles , in galactic coordinates , for all 709 objects .
error bars designate poisson uncertainties for each bin .
while the overall distribution is generally consistent with random , there is a small excess of objects at 90 and 270 orientations , parallel to the plane .
figure [ sizes ] shows a histogram of the base 10 logarithm of the standoff distance , @xmath41 , in arcseconds , describing the separation of the central star from the apsis of the arc - shaped nebula .
the peak of the distribution in the range 512 indicates that the vast majority of the objects are compact , sufficiently small that the angular resolution of @xmath15 is required to discern the arc - like shape of the nebula .
the distribution declines smoothly from 12 to about 50 , beyond which a small number of very large arcs with sizes up to 500 populate the distribution more or less uniformly . on the smaller end of the distribution
the number of objects drops steeply from sizes of 5 to the resolution limit if the irac instrument at about 2 .
figure [ avr0 ] plots the angular size of the objects , @xmath25 , versus extinction toward the star , @xmath28 , as inferred from the @xmath424.5 @xmath0 m color excess , assuming a standard extinction law with @xmath28= 9@xmath20@xmath43 .
a weak correlation is evident , with the largest nebulae being the least extinguished and the smallest nebulae having preferentially larger extinctions . if extinction of the starlight were a proxy for distance , we might expect that the most extinguished objects would also have the smallest angular sizes , but no strong trends are evident .
figure [ environ ] depicts the fraction of the bowshock candidate sample that lies in apparently isolated environments ( i ; 70.4% ) , facing regions ( fh ; 14.5% ) , facing bright - rimmed clouds ( fb ; 8.2% ) , and located within regions ( h ; 6.9% ) .
many objects designated as fb also lie within regions , so fb may entail h ; fb and h together make up 15.3% of the sample .
isolated objects ( 499 ) make up more than two - thirds of the sample .
however , this designation does not necessarily mean that the object is physically isolated from other potential astrophysical influences .
in the complex regions near the galactic plane , most candidates are seen in projection near dark clouds , star forming regions , and other prominent features .
rather , the isolated designation simply means that the object does not obviously appear to lie within or to directly face a visible region , star or bright - rimmed cloud . as such , these are the best candidates for being genuine runaway stars with space velocities sufficient to sweep up ambient material into an infrared - emitting bowshock nebula .
about 14.6% ( 103 objects ) of the sample exhibits a bowshock orientation that faces a prominent region within @xmath4410 . in such close proximity , outflows from a hot ionized bubble could plausibly interact with the wind of the ob star , producing an `` in - situ '' type bowshock even if the star does not have a high space velocity .
the association of a bowshock nebulae with a particular region is necessarily speculative , but the relative orientation occurs frequently enough in our sample that the physical nature of this class of objects is highly suggestive . eight percent ( 58 objects ) of the sample faces a bright - rimmed cloud within several arcminutes .
bright - rimmed features represent the photo - excited surfaces of molecular clouds emitting strongly in the 8.0 @xmath0 m bandpass by virtue of the solid state features of pahs .
they occur commonly at the peripheries of regions where massive stars create cavities or bubbles in the surrounding gas , e.g , the collection of interstellar bubbles cataloged by @xcite .
figure [ examples.fb ] shows six objects that face bright - rimmed clouds , with the color scheme as in figure [ prototypes ] .
frequently , this class of objects is located _ within _ an region .
physically , this class of objects may arise when an ob star s wind interacts with a photo - ablated flow from the molecular cloud boundary produced by the uv photons from the same star .
the phenomenon may be similar to the stationary shocks visible as ionized arcs at the interface between the wind from @xmath9 ori c and the proplyds in the orion nebula @xcite .
figure 3 of @xcite conveniently illustrates the location of one object , proplyd 180 - 331 , that hosts an arc facing the ionization front in the omc1 molecular cloud rather than the facing the star @xmath9 ori c. given that the phenomena operating in our subsample of ir - selected bowshock may be similar , we term this subclass of of in - situ bowshock nebulae `` photo - evaporative flow '' ( pef ) bowshocks .
however , in orion , the arcs are very thin features visible primarily in @xmath45 and , whereas our sample is prominent in the mid - ir dust continuum and the ionization structure is not presently known .
one prediction following from this interpretation is that the pef bowshock candidates , designated here as fb , do not have high space velocities relative to their local environment .
the remaining 7.0% ( 49 objects ) appear to lie within regions or interstellar bubbles , as outlined by either bright - rimmed clouds or by pillars of photo - excited gas .
while they do not obviously face a bright - rimmed cloud at the periphery of the region , they may be produced where a stellar wind interacts with a high - velocity flow of material , including champagne flows .
even so , these objects may overlap with the fb subsample which are frequently found within regions .
following the analysis of @xcite , we collected proper motion measurements for the stellar counterparts to each bowshock nebula . in this analysis
we use proper motion data from the hipparcos @xcite , ucac4 @xcite , and the usno b2 @xcite catalogs , in order of preference . of the 709candidates , only 40% ( 286 )
have proper motion data of any kind , and of these only 139 have sufficiently small uncertainties that the proper motion position angles are known to better than 45 . nevertheless , we computed the measured the angle between the nominal orientation angle of the putative bowshock nebula ( column 10 of table [ bigtable.tab ] ) and the proper motion vector .
figure [ deltapa ] displays a histogram of this distribution in 45 bins between @xmath26180 and 180 .
error bars depict the uncertainties , taken to be the square root of the number of objects in each bin .
a clear excess in the two bins on either side of zero indicates a preference for proper motions to coincide with the expected direction inferred from the infrared morphology .
this suggests that at least 25% of the objects are good candidates for being runaway stars with high space velocities in the direction of apsis of the infrared arc .
however , the fraction of sources in this plot having relative position angles exceeding @xmath46 is large compared to the sample studied by @xcite . because our sample was selected on the basis of ir morphology rather than on the basis of stars known to have high proper motion ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , it is not surprising that our sample displays a broader distribution of relative position angles .
the broad distribution of relative position angles may indicate that factors other than a star s peculiar motion could produce relative velocities between a star and the surrounding interstellar material that generate a bowshock nebula .
section [ env.sec ] above discusses physical scenarios that could explain these objects .
we have compiled the largest collection to date of candidate stellar bowshock nebulae selected on the basis of mid - ir morphology at 24 or 22 @xmath0 m from @xmath15 and @xmath12 sky surveys of the galactic plane .
the vast majority 660 of the 709 objects are identified here for the first time .
a minority of objects ( @xmath119% ) appear at 8 @xmath0 m or shorter wavelengths as well , indicating a population of very small hot dust grains within the swept - up nebula , or possibly a pah contribution .
while the infrared images alone do not demonstrate that the tabulated objects are necessarily shocks , the preponderance of ancillary evidence , including proper motions and spectral classifications of the host stars , strongly suggests that this is the best interpretation for the majority of objects . on further investigation
, some of the objects may turn out to be other phenomena , such as asymmetric dust shells around evolved stars .
the distribution of these objects on the sky is tightly confined to the galactic mid - plane , consistent with their production by massive stars which have strong stellar winds .
proper motions of the central stars , where known , indicate a clear excess of objects having velocity vectors aligned with the symmetry axis of the infrared nebula .
this is consistent with their production by a population of massive runaway stars . while the majority of the candidate bowshock nebulae and their central stars lie in relatively isolated environments , a substantial subset ( @xmath120% ) either face giant regions or face bright - rimmed clouds
these may , respectively , constitute two classes of `` in - situ '' bowshocks where either an external flow overruns the star or `` photoevaporative flow '' bowshocks where a stellar wind interacts with material evaporating from a nearby molecular cloud skin .
the correlation of bowshock orientations on angular scales of @xmath610 provides evidence that local environmental phenomena , such as outflows from star forming regions , produce the relative motions between a nearby massive star and the ism that generate the bowshock nebulae .
table [ bigtable.tab ] and the appendix of 709stellar bowshock candidates may be used to further investigate the origins of runaway stars , phenomena associated with large - scale outflows of gas from star forming regions , and the physics of massive star winds .
additional scrutiny of existing infrared surveys could be expected to find a few additional bowshock candidates , especially at higher galactic latitudes not covered by our search .
however , we expect the yield to be small as the areal density of bowshocks within the @xmath15 plane surveys drops rapidly with latitude . in a forthcoming series of papers we will present an analysis of the characteristics of this comprehensive bowshocks sample , including their spectral types , binarity , distances , spectral energy distributions , and local environments . our team is grateful to support from the national science foundation through grant ast-1412845 , ast-1411851 , and reu grant ast-1063146 , as well as nasa through grant nnx14ar35a .
we thank our anonymous reviewer for comments that helped clarify aspects of this manuscript .
acreman , d. m. , stevens , i. r. , & harries , t. j. 2016 , , 456 , 136 bally , j. , sutherland , r. s. , devine , d. , & johnstone , d. 1998 , , 116 , 293 benjamin , r. , et al . 2003 , pasp 115 , 953 blaauw , a. 1961 , bulletin of the astronomical institutes of the netherlands , 15 , 265 blaauw , a. 1993 , in massive stars : their lives in the interstellar medium , eds .
j. p. cassinelli , & e. b. churchwell , asp conf .
ser . , 35 , 207 bonnarel , f. , fernique , p. , bienaym , o. , et al .
2000 , , 143 , 33 brown , w. r. , geller , m. j. , kenyon , s. j. , kurtz , m. j. 2005 , apj , 622 , l33 brown , w. r. , geller , m. j. , kenyon , s. j. 2014 , apj , 787 , 89 carey , s. j. , noriega - crespo , a. , mizuno , d. r. , et al .
2009 , , 121 , 76 chick , w. t. , kobulnicky , h. a. , schurhammer , d. , et al .
2016 churchwell , e. , povich , m. s. , allen , d. , et al .
2006 , , 649 , 759 carey , s. , etal .
2008 , spitzer proposal i d # 50398 comern , f. , & kaper , l. 1998 , a&a , 338 , 273 comern , f. , & pasquali , a. 2007 , , 467 , l23 cox , n. l. j.,kerschbaum , f. , van marle , a. j. , decin , l. , et al .
2012 , a&a , 537 , 35 de wit w. j. , testi , l. , palla , f. , zinnecker , h. 2005 , a&a , 437 , 247 fazio , g. g. , hora , j. l. , allen , l. e. , et al .
2004 , , 154 , 10 fryer , c. , burrows , a. , benz , w. 1998 , apj , 496 , 333 fuentes , c. i. , stanek , k. z. gaudi , b. s. , et al .
2006 , apj , 636 , l37 garca , b. , & mermilliod , j. c. 2001 , a&a , 368 , 122 garca - arredondo , f. , henney , w. j. , & arthur , s. j. 2001 , , 561 , 830 gspr , a. , su , k. y. l. , rieke , g. h. , et al .
2008 , , 672 , 974 gies , d. r. & bolton , c. t. 1986 , , 419 , 61 gull , t. r. & sofia , s. 1979 , , 230 , 782 gvaramadze , v. v. , & bomans , d. j. 2008 , , 490 , 1071 gvaramadze , v. v. 2009 , mnras , 395 , l85 gvaramadze , v. v. , kroupa , p. , & pflamm - altenburg , j. 2010 , a&a , 519 , 33 gvaramadze , v. v. , pflamm - altenburg , j. , & kroupa , p. 2011a , a&a , 525 , 17 gvaramadze , v. v. , rser , s. , scholz , r .- d . ,
schilbach , e. 2011b , a&a , 529 , 14 gvaramadze , v. v. , kniazev , a. y. , kroupa , p. , oh , s. 2011c , a&a , 535 , a29 gvaramadze , v. v. , kniazev , a. y. , chen a .-
n . , schnurr , o. 2013 , mnras , 430 , l20 gvaramadze , v. v. , menten , k. m. , kniazev , a. y. , et al . 2014a , mnras , 437 , 843 gvaramadze , v. v. , miroshnichenko , a. s. , castro , n. , langer , n. , zharikov , s. v. 2014b , mnras , 437 , 2761 hills , j. g. 1988 , nature , 331 , 687 hills , j. g. 1991 , aj , 102 , 704 hoogerwerf , r. , de bruijne , h. j. , & de zeeuw , p. t. 2000 , , 544 , l133 hoogerwerf , r. , de bruijne , h. j. , & de zeeuw , p. t. 2001 , a&a , 365 , 49 hora , j. , et al . 2007 ,
spitzer proposal i d # 40184 kenyon s. j. , bromley , b. c. , geller , m. j. , brown , w. r. 2008 , apj , 680 , 312 kiminki , d. c. , & kobulnicky , h. a. 2012 , , 751 , 4 kobulnicky , h. a , gilbert , i. j. , & kiminki , d. c. 2010 , , 710 , 549 kobulnicky , h. a. , kiminki , d. c. , lundquist , m. j. , et al . 2014 , , 213 , 34 kroupa , p. , aarseth , s. , & hurley , j. 2001 , mnras , 321 , 699 leonard , p.j.t .
1991 , , 101 , 562 liu , z .- w .
, tauris t. m. , rpke , f. k. , et al .
2015 , a&a , 584 , 11 mac low , m .-
m . , klein , r. i. , mckee , c. f. , stone , j. , norman , m. l. 1991 , , 369 , 395 majewski , s. r. , zasowski , g. , & nidever , d. l. 2011 , , 739 , 25 meyer , d. m .- a . , mackey , j. , langer , n. , et al .
2014 , mnras , 444 , 2754 meyer , d. m .- a . , van marle , a .- j . ,
kuiper , r. , kley , w. 2016 , mnras , 459 , 1146 monet , d. g. , levine , s. e. , canzian , b. , et al .
2003 , , 125 , 984 neugebauer , g. , habing , h. j. , van duinen , r. , et al .
1984 , , 278 , l1 noriega - crespo , a. , van buren , d. , & dgani , r. 1997 , , 113 , 780 peri , c. s. , benaglia , p. , brookes , d. p. , stevens , i. r. , & isequilla , n. l. 2012 , , 538 , a108 peri , c. s. , benaglia , p. , & isequilla , n. l. 2015 , , 578 , a45 perryman , m. a. c. , lindegren , l. , kovalevsky , j. , et al .
1997 , , 323 , pflamm - altenburg , j. & kroupa p. 2010 , 404 , 1564 pittard , j. m. 2009 , , 396 , 1743 pittard , j. m. , & dougherty , s. m. 2006 , , 372 , 801 poveda , a. , ruiz , j. , & allen , c. 1967 , bol .
. tonantzintla y tacubaya , 4 , 86 povich , m. s. , benjamin , r. a. , whitney , b. a. , et al . 2008 , , 689 , 242 - 248 rieke , g. h. , young , e. t. , engelbracht , c. w. , et al .
2004 , , 154 , 25 sana , h. , de mink , s. e. , de koter , a. , et al .
2012 , science , 337 , 444 sexton , r. o. , povich , m. s. , smith , n. , et al .
2015 , , 446 , 1047 skrutskie , m. f. , cutri , r. m. , stiening , r. , et al .
2006 , , 131 , 1163 spitzer , l. jr . ,
& mathieu , r. d. 1980 , , 241 , 618 stevens , i. r. , blondin , j. m. , & pollock , a. m. t. 1992 , , 386 , 265 stone , r. c. 1991 , aj , 102 , 333 tauris , t. m. 2015 , mnras , 448 , l6 tetzlaff , n. , neuhauser , r. , & hohle , m. m. 2010 , mnras , 410 , 190 usov , v. v. 1992 , , 389 , 635 van buren , d. , & mccray , r. 1988 , , 329 , l93 van buren , d. , mac low , m .-
m . ; wood , d. o. s. ; churchwell , e. 1990 , , 353 , 570 van buren , d. & mac low , m .-
1992 , , 394 , 534 van buren , d. , noriega - crespo , a. , & dgani , r. 1995 , , 110 , 2914 wang , z. , kaplan , d. l. , slane , p. , morrell , n. , kaspi , v. m. 2013 , , 769 , 122 werner , m. w. , roellig , t. l. , low , f. j. , et al .
2004 , , 154 , 1 wilkin , f. p. 1996
, , 459 , l31 wilkin , f. p. 2000 , , 532 , 400
wright , e. l. , eisenhardt , p. r. m. , mainzer , a. k. , et al .
2010 , , 140 , 1868 - 1881 yu , q. & tremaine , s. 2003 , 599 , 1129 zacharias , n. , finch , c. t. , girard , t. m. , et al .
2013 , , 145 , 44 zwicky , f. 1957 , @xmath47 ( berlin : springer ) , p. 258
cccccccccccccc 1 & [email protected] & 17:48:07.70 & @xmath2629:07:55.5 & t & & y & c & 26.4 & 27 & 8.97 & 8.72 & 0.22 & i + 2 & [email protected] & 17:50:27.59 & @xmath2629:12:46.8 & t & & n & c & 6.8 & 145 & 10.57 & 10.31 & 0.23 & i + 3 & [email protected] & 17:48:41.78 & @xmath2628:06:37.8 & t & & y & & 6.2 & 200 & 11.14 & 9.8 & 1.22 & h + 4 & [email protected] & 17:48:53.40 & @xmath2627:53:59.7 & t & & n & c & 10.2 & 215 & 12.28 & 11.22 & 0.96 & fh + 5 & [email protected] & 17:53:56.10 & @xmath2625:56:50.3 & t & & n & & 8.6 & 135 & 10.69 & 9.69 & 0.91 & i + 6 & g003.7391 + 00.1425 & 17:53:43.26 & @xmath2625:39:19.2 & t & & n & & 3.9 & 20 & 11.48 & 9.41 & 1.89 & i + 7 & [email protected] & 17:58:30.64 & @xmath2626:09:49.1 & t & & y & & 12.6 & 25 & 10.22 & 9.87 & 0.31 & fb + 8 & g004.3087 + 00.2222 & 17:54:41.61 & @xmath2625:07:25.6 & t & & n & & 10.1 & 260 & 13.94 & 12.1 & 1.68 & i + 9 & [email protected] & 17:57:57.62 & @xmath2625:03:53.6 & t & & n & c & 10.1 & 125 & 13.3 & 11.77 & 1.39 & i + 10 & [email protected] & 18:00:17.51 & @xmath2625:14:14.7 & t & hd314937 & n & & 3.1 & 140 & 9.36 & 9.31 & 0.03 & i + 11 & g005.5941 + 00.7335 & 17:55:36.25 & @xmath2623:45:21.8 & t & & n & & 7.6 & 350 & 9.78 & 9.22 & 0.50 & i + 12 & [email protected] & 18:01:06.15 & @xmath2624:21:34.4 & t & & n & & 20.8 & 180 & 11.56 & 9.11 & 2.24 & i + 13 & g006.2812 + 23.5877 & 16:37:09.54 & @xmath2610:34:01.5 & v88 & zeta oph & - & & 29 & 30 & 2.59 & -99 & -99 & i + 14 & [email protected] & 18:00:40.59 & @xmath2623:36:50.7 & t & & n & c & 24 & 210 & 13.07 & 12.16 & 0.82 & h + 15 & [email protected] & 18:00:44.92 & @xmath2623:33:06.3 & t & & n & & 5.5 & 110 & 9.32 & 9.11 & 0.18 & i + 16 & g006.8933 + 00.0743 & 18:00:55.27 & @xmath2622:57:36.9 & t & & n & c & 29 & 65 & 8.73 & 8.36 & 0.33 & i + 17 & [email protected] & 18:03:33.49 & @xmath2622:34:38.3 & t & & n & & 4.1 & 350 & 9.06 & 8.11 & 0.86 & i + 18 & g008.3690 + 00.0239 & 18:04:15.53 & @xmath2621:42:05.3 & t & & n & c & 27 & 50 & 9.07 & 8.3 & 0.69 & i + 19 & g009.0177 + 00.1410 & 18:05:11.21 & @xmath2621:04:43.3 & t & & n & & 9.1 & 290 & 8.85 & 8.41 & 0.39 & i + 20 & [email protected] & 18:07:51.86 & @xmath2620:39:48.8 & t & & n & & 8.3 & 335 & 10.92 & 9.69 & 1.12 & i
this electronic - only appendix contains an atlas of three - color images for each of the 709 bowshock nebula candidates . for the majority of the candidates that have @xmath15 data ,
the color scheme is 24/8.0/4.5 @xmath0 m in red / green / blue . for objects outside the coverage of the @xmath15 surveys , we use the @xmath12 22/12/4.6 bands in red / green / blue . in a few cases where the @xmath15 24 @xmath0 m data are not available
, we substitute the @xmath12 22 @xmath0 m data . images are 2 square unless a larger field of view is required to show the entire nebulae or pertinent surrounding features .
white arrows on some images depict the direction and magnitude of central star s proper motion , where known . | we identify 709 arc - shaped mid - infrared nebula in 24 @xmath0 m _ spitzer space telescope _ or 22 @xmath0 m _ wide field infrared explorer
_ surveys of the galactic plane as probable dusty interstellar bowshocks powered by early - type stars .
about 20% are visible at 8 @xmath0 m or shorter mid - infrared wavelengths as well . the vast majority ( 660 ) have no previous identification in the literature .
these extended infrared sources are strongly concentrated near galactic mid - plane with an angular scale height of @xmath10.6 .
all host a symmetrically placed star implicated as the source of a stellar wind sweeping up interstellar material .
these are candidate `` runaway '' stars potentially having high velocities in the reference frame of the local medium . among the 286 objects with measured proper motions
, we find an unambiguous excess having velocity vectors aligned with the infrared morphology kinematic evidence that many of these are `` runaway '' stars with large peculiar motions responsible for the bowshock signature .
we discuss a population of `` in - situ '' bowshocks ( @xmath1103 objects ) that face giant regions where the relative motions between the star and ism may be caused by bulk outflows from an overpressured bubble .
we also identify @xmath158 objects that face 8 @xmath0 m bright - rimmed clouds and apparently constitute a sub - class of in - situ bowshocks where the stellar wind interacts with a photo - evaporative flow from an eroding molecular cloud interface ( i.e. , `` pef bowshocks '' ) .
orientations of the arcuate nebulae exhibit a correlation over small angular scales , indicating that external influences such as regions are responsible for producing some bowshock nebulae .
however , the vast majority of this sample appear to be isolated ( 499objects ) from obvious external influences . |
a semialgebraic set is a subset of @xmath0 which is a solution set of a system of polynomial equations and inequalities .
computation with semialgebraic sets is one of the core subjects in computer algebra and real algebraic geometry .
a variety of algorithms have been developed for real system solving , satisfiability checking , quantifier elimination , optimization and other basic problems concerning semialgebraic sets @xcite .
every semialgebraic set can be represented as a finite union of disjoint cells bounded by graphs of algebraic functions . the cylindrical algebraic decomposition ( cad ) algorithm @xcite can be used to compute a cell decomposition of any semialgebraic set presented by a quantified system of polynomial equations and inequalities .
an alternative method of computing cell decompositions is given in @xcite .
cell decompositions computed by the cad algorithm can be represented directly @xcite as cylindrical algebraic formulas ( caf ; see the next section for a precise definition ) .
a caf representation of a semialgebraic set @xmath1 can be used to decide whether @xmath1 is nonempty , to find the minimal and maximal values of the first coordinate of elements of @xmath1 , to generate an arbitrary element of @xmath1 , to find a graphical representation of @xmath1 , to compute the volume of @xmath1 , or to compute multidimensional integrals over @xmath1 ( see @xcite ) .
the cad algorithm takes a system of polynomial equations and inequalities and constructs a cell decomposition of its solution set .
the algorithm consists of two phases .
the projection phase finds a set of polynomials whose roots are sufficient to describe the cell boundaries .
the lifting phase constructs a cell decomposition , one dimension at a time , subdividing cells at all roots of the projection polynomials .
however , some of these subdivisions may be unnecessary , either because of the geometry of the roots or because of the boolean structure of the input system . in this paper
we propose an algorithm which combines the two phases .
it starts with a sample point and constructs a cell containing the point on which the input system has a constant truth value .
projection polynomials used to construct the cell are selected based on the structure of the system at the sample point .
such a local projection set can often be much smaller than the global projection set used by the cad algorithm .
the idea to use such locally valid projections was first introduced in @xcite , in an algorithm to decide the satisfiability of systems of real polynomial equations and inequalities .
it was also used in @xcite , in an algorithm to construct a single open cell from a cylindrical algebraic decomposition .
[ exa : mainexample]find a cylindrical algebraic decomposition of the solution set of @xmath2 , where @xmath3 , @xmath4 , and @xmath5 .
the solution set of @xmath6 is equal to the union of the open ellipse @xmath7 and the intersection of the closed disk @xmath8 and the set @xmath9 bounded by a lissajous curve .
as can be seen in the picture , the set is equal to the open ellipse @xmath7 .
the cad algorithm uses a projection set consisting of the discriminants and the pairwise resultants of @xmath10 , @xmath11 , and @xmath12 .
it computes a cell decomposition of the solution set of @xmath6 by constructing @xmath13 cells such that all @xmath10 , @xmath11 , and @xmath12 have a constant sign on each cell .
note however , that a cell decomposition of the solution set of @xmath6 can be obtained by considering the following @xmath14 cells .
on each cell only some of @xmath10 , @xmath11 , and @xmath12 have a constant sign , but those signs are sufficient to determine the truth value of @xmath6 . 1 .
@xmath6 is @xmath15 on @xmath16 because @xmath7 .
2 . @xmath6 is @xmath17 on @xmath18 and on @xmath19 because @xmath20 .
3 . @xmath6 is @xmath17 on @xmath21 and on @xmath22 because @xmath23 .
4 . @xmath6 is @xmath17 on @xmath24 and on @xmath25 because @xmath20 .
@xmath6 is @xmath17 on @xmath26 and on @xmath27 because @xmath20 .
@xmath6 is @xmath17 on @xmath28 because @xmath29 .
@xmath6 is @xmath17 on @xmath30 and on @xmath31 because @xmath20 .
@xmath6 is @xmath17 on @xmath32 because @xmath29 .
determining the cell bounds for the cell stack @xmath33-@xmath34 requires computation of roots of @xmath35 , @xmath36 , and @xmath37 in @xmath38 and roots of @xmath39 and @xmath40 in @xmath41 .
determining the cell bounds for the cells @xmath42 requires computation of roots of @xmath35 and @xmath36 in @xmath38 and roots of @xmath43 , @xmath44 , @xmath45 and @xmath46 in @xmath41 .
determining the cell bounds for the cell stacks @xmath47-@xmath48 and @xmath49-@xmath50 requires computation of roots of @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 and @xmath56 in @xmath41 .
polynomial @xmath12 is not used to compute any of the projections and its roots in @xmath41 are computed only for two values of @xmath38 .
the algorithm we propose in this paper computes a cell decomposition of the solution set of @xmath6 by constructing the @xmath14 cells given in @xmath33-@xmath50 .
details of the computation for this example are given in section [ sub : example ] .
find a cylindrical algebraic decomposition of the solution set of @xmath57 in the variable order @xmath58 . in this example
the system is not well - oriented , hence the cad algorithm needs to use hong s projection operator for the first three projections .
however , the additional projection polynomials are necessary only for the cells on which a mccallum s projection polynomial vanishes identically . for most cells
local projection can be computed using mccallum s projection operator , and for the few cells on which a mccallum s projection polynomial vanishes identically local projection needs to use some , but usually not all , polynomials from hong s projection operator .
the algorithm lpcad we propose in this paper computes a cell decomposition of the solution set of @xmath6 by constructing @xmath59 cells in @xmath60 seconds of cpu time .
the cad algorithm did not finish the computation in @xmath61 hours .
a version of lpcad using only local projections based on hong s projection operator constructs @xmath62 cells and takes @xmath63 seconds of cpu time .
a _ system of polynomial equations and inequalities _ in variables @xmath64 is a formula@xmath65 where @xmath66 $ ] , and each @xmath67 is one of @xmath68 or @xmath69 .
a subset of @xmath0 is _ semialgebraic _ if it is a solution set of a system of polynomial equations and inequalities .
a _ quantified system of real polynomial equations and inequalities _ in free variables @xmath64 and quantified variables @xmath70 is a formula @xmath71 where @xmath72 is @xmath73 or @xmath74 , and @xmath6 is a system of real polynomial equations and inequalities in @xmath75 . by tarski s theorem ( see @xcite ) ,
solution sets of quantified systems of real polynomial equations and inequalities are semialgebraic . for @xmath76 ,
let @xmath77 denote a @xmath78-tuple @xmath79 of real numbers and let @xmath80 denote a @xmath78-tuple @xmath81 of variables .
every semialgebraic set can be represented as a finite union of disjoint _ cells _ ( see @xcite ) , defined recursively as follows . 1 .
a cell in @xmath82 is a point or an open interval .
a cell in @xmath83 has one of the two forms@xmath84 where @xmath85 is a cell in @xmath86 , @xmath87 is a continuous algebraic function , and @xmath88 and @xmath89 are continuous algebraic functions , @xmath90 , or @xmath91 , and @xmath92 on @xmath85 .
a finite collection @xmath93 of cells in @xmath0 is _ cylindrically arranged _ if for any @xmath94 and @xmath95 the projections of @xmath96 and @xmath97 on @xmath86 are either disjoint or identical . given a semialgebraic set presented by a quantified system of polynomial equations and inequalities , the cad algorithm can be used to decompose the set into a cylindrically arranged finite collection of cells .
the collection of cells is represented by a cylindrical algebraic formula ( caf ) .
a caf describes each cell by giving explicit algebraic function bounds and the boolean structure of a caf reflects the cylindrical arrangement of cells . before we give a formal definition of a caf ,
let us first introduce some terminology .
let @xmath76 and let @xmath98 , where @xmath99 $ ] .
a _ real algebraic function _ given by the _ defining polynomial _ @xmath100 and a _ root number _
@xmath101 is the function@xmath102 where @xmath103 is the @xmath104-th real root of @xmath105 $ ] .
the function is defined for those values of @xmath77 for which @xmath106 has at least @xmath104 real roots .
the real roots are ordered by the increasing value and counted with multiplicities .
a real algebraic number @xmath107 given by a _ defining polynomial _
@xmath108 $ ] and a _ root number _
@xmath104 is the @xmath104-th real root of @xmath100 .
see @xcite for more details on how algebraic numbers and functions can be implemented in a computer algebra system .
let @xmath109 be a connected subset of @xmath86 .
@xmath110 is _ regular _ on _ @xmath109 _ if it is continuous on @xmath109 , @xmath111 for all @xmath112 , and there exist _ _
@xmath113 such that for any @xmath112 @xmath103 is a root of @xmath106 of multiplicity @xmath114 .
@xmath100 is _ degree - invariant _ on @xmath109 if there exist _
_ @xmath115 such that if @xmath116 for all @xmath112 .
a set @xmath117 of polynomials is _ delineable _ on @xmath109 if all elements of @xmath118 are degree - invariant on @xmath109 and for @xmath119@xmath120 where @xmath121 are disjoint regular real algebraic functions and for @xmath122 @xmath123 and @xmath124 are either disjoint or equal .
functions @xmath125 are _ root functions of @xmath126 over @xmath109_. a set @xmath117 of polynomials is _ analytic delineable _ on a connected analytic submanifold @xmath109 of @xmath86 if @xmath118 is delineable on @xmath109 and the root functions of elements of @xmath118 over @xmath109 are analytic .
let @xmath118 be delineable on @xmath109 , let @xmath127 be all root functions of elements of @xmath118 over @xmath109 , and let @xmath128 and @xmath129 . for @xmath130 , the @xmath131-th _
@xmath118-section over @xmath109 _ is the set@xmath132 for @xmath133 , the @xmath131-th _
@xmath118-sector over @xmath109 _ is the set@xmath134 a formula @xmath135 is an _ algebraic constraint _ with
_ bounds _ @xmath136 if it is a level-@xmath78 equational or inequality constraint with @xmath137 defined as follows .
_ _ 1 .
_ a level_-@xmath138 _ equational constraint _ has the form @xmath139 , where @xmath87 is a real algebraic number , and @xmath140 .
a level_-@xmath138 _ inequality constraint _ has the form @xmath141 , where @xmath88 and @xmath89 are real algebraic numbers , @xmath90 , or @xmath91 , and @xmath142 .
a level_-@xmath143 _ equational constraint _ has the form @xmath144 , where @xmath87 is a real algebraic function , and @xmath140 .
a level_-@xmath143 _ inequality constraint _ has the form @xmath145 , where @xmath88 and @xmath89 are real algebraic functions , @xmath90 , or @xmath91 , and @xmath142 .
a level-@xmath143 algebraic constraint @xmath135 is _ regular _ on a connected set @xmath146 if all elements of @xmath136 are regular on @xmath109 and , if @xmath135 is an inequality constraint , @xmath92 on @xmath109 .
an _ atomic cylindrical algebraic formula ( caf ) _ @xmath135 in @xmath147 has the form @xmath148 , where @xmath149 is a level-@xmath78 algebraic constraint for @xmath137 and @xmath150 is regular on the solution set of @xmath151 for @xmath152 . _
level-@xmath78 cylindrical subformulas _ are defined recursively as follows 1 .
a level-@xmath153 cylindrical subformula is a disjunction of level-@xmath153 algebraic constraints .
2 . a level-@xmath78 cylindrical subformula , with @xmath152 ,
has the form@xmath154 where @xmath155 are level-@xmath78 algebraic constraints and @xmath156 are level-@xmath143 cylindrical subformulas .
a _ cylindrical algebraic formula ( caf ) _ is a level-@xmath138 cylindrical subformula @xmath135 such that distributing conjunction over disjunction in @xmath135 gives @xmath157 where each @xmath155 is an atomic caf . given a quantified system of real polynomial equations and inequalities the cad algorithm @xcite returns a caf representation of its solution set . the following formula @xmath158 is a caf representation of the closed unit ball.@xmath159 where @xmath160
in this section we describe an algorithm for computing a caf representation of the solution set of a system of polynomial equations and inequalities .
the algorithm uses local projections computed separately for each cell . for simplicity
we assume that the system is not quantified .
the algorithm can be extended to quantified systems following the ideas of @xcite .
the algorithm in its version given here does not take advantage of equational constraints .
the use of equational constraints will be described in the full version of the paper .
the main , recursive , algorithm used for cad construction is algorithm [ alg : lpcad ] .
let us sketch the algorithm here , a detailed description is given later in this section . the input is _ _ a system @xmath161 _ _ of polynomial equations and inequalities and a point @xmath162 with @xmath163 .
the algorithm finds a level-@xmath143 cylindrical subformula @xmath135 and a set of polynomials _
@xmath164 $ ] _ such that for any cell @xmath146 containing @xmath77 on which all elements of @xmath165 have constant signs _ _ @xmath166 the formula @xmath135 can be interpreted as a description of the solution set of @xmath6 as a finite collection of cylindrically arranged cells in @xmath167 , parametrized by the values of @xmath81 .
the description is valid locally to @xmath77 , where the meaning of `` locally '' is determined by @xmath165 .
the approach is to find algebraic constraints @xmath168 and cylindrical subformulas @xmath169 such that the solution sets of @xmath170 form a decomposition of @xmath82 and @xmath171 describes the solution set of @xmath6 locally to @xmath172 . to find @xmath173 s , @xmath174 s , and @xmath165 we start with a stack containing the interval @xmath175 and until the stack is emptied execute the following steps .
we take an interval @xmath176 off stack and pick @xmath177 .
if evaluating the @xmath143-variate polynomials in @xmath6 at @xmath178 suffices to establish the truth value of @xmath6 , let @xmath179 be a set of @xmath143-variate polynomials in @xmath6 sufficient to establish the truth value of @xmath6 and let @xmath174 be the truth value .
otherwise , let @xmath174 and @xmath179 be , respectively , the formula and the set of polynomials returned by algorithm [ alg : lpcad ] applied to @xmath6 and @xmath178 .
we use projection to compute a set _ @xmath180 $ ] _ such that @xmath179 is delineable on any cell containing @xmath77 on which all elements of @xmath118 have constant signs and we add the elements of @xmath118 to @xmath165 .
let @xmath181 be the interval containing @xmath182 bounded by the nearest roots of elements of @xmath179 and let @xmath173 be the constraint on @xmath183 whose bounds are the corresponding algebraic functions .
note that if @xmath179 is delineable on a cell @xmath109 containing @xmath77 then the elements of @xmath179 have constant signs on @xmath184 and hence @xmath174 is equivalent to @xmath6 on @xmath93 .
we add @xmath173 and @xmath174 to the list of @xmath173 s , @xmath174 s , and , if @xmath185 is nonempty , we add the components of @xmath185 to stack . when the stack is empty we use projection to compute a set _
@xmath180 $ ] _ such the set of polynomials whose roots appear as bounds in @xmath173 s are delineable on any cell containing @xmath77 on which all elements of @xmath118 have constant signs and we add the elements of @xmath118 to @xmath165 . as required
, the formula @xmath186 is equivalent to @xmath6 on any cell containing @xmath77 on which all elements of @xmath165 have constant signs . to compute a caf representation of the solution set of @xmath6 we call algorithm [ alg : lpcad ] with @xmath187 .
we will use the following notations . 1 . for a finite set of polynomials @xmath179 , let @xmath188 denote the set of irreducible factors of the elements of @xmath179 .
2 . let @xmath189 denote the irreducible elements of @xmath190\setminus\mathbb{r}[x_{1},\ldots , x_{k-1}]$ ] .
3 . for a set @xmath191 and @xmath95
let @xmath192 denote the projection of @xmath1 on @xmath86 .
in this section we assume that all polynomials have coefficients in a fixed computable subfield @xmath193 , irreducibility is understood to be in the ring of polynomials with coefficients in @xmath194 , irreducible factors are always content - free and chosen in a canonical way , and finite sets of polynomials are always ordered according to a fixed linear ordering in the set of all polynomials with coefficients in @xmath194 . in our implementation @xmath195 . whenever we write _ @xmath196 _ with @xmath197 we include the possibility of @xmath198 , the only element of @xmath199 .
let @xmath200 $ ] be a finite set of polynomials and let @xmath201 , where @xmath202 .
let @xmath203 be such that @xmath204 is a finite subset of @xmath189 and @xmath205 for @xmath137 .
@xmath118 is a _ local projection sequence _ for @xmath179 at @xmath206 iff , for any @xmath152 and any cell @xmath146 , if @xmath207 and all elements of @xmath208 for @xmath209 have constant signs on @xmath210 then the set of elements of @xmath211 that are not identically zero on @xmath212 is delineable over @xmath109 .
to compute local projections we use the following two projection procedures , derived , respectively , from mccallum s projection operator @xcite and hong s projection operator @xcite .
[ alg : lprojmc](lprojmc ) + input : _
_ @xmath213 _ and _ @xmath162 _ _ , where @xmath76 .
_ _ + _ finite set _
@xmath214 $ ] .
1 . put @xmath215 and compute @xmath216 .
2 . for @xmath119
let @xmath217 .
put @xmath218 .
2 . if @xmath219 and @xmath220 put @xmath221 and continue the loop .
3 . if @xmath219 , @xmath222 , and none of @xmath223 is a nonzero constant , put @xmath224 , where @xmath225 is maximal such that @xmath226 .
4 . put @xmath227 .
5 . if @xmath228 then put @xmath229 3 .
return @xmath230 . in the next algorithm
we use the following notation .
let @xmath231[x_{k+1}]$ ] , @xmath232 , and @xmath233 if for some @xmath234 , @xmath235 and @xmath236 , then @xmath237 .
otherwise @xmath238
[ alg : lprojh](lprojh ) + input : _ _ @xmath213 _ and _ @xmath162 _ _ , where @xmath76 . _ _ + _ finite set _
@xmath214 $ ] .
1 . put @xmath215 and compute @xmath216 .
2 . for @xmath119
let @xmath217 .
put @xmath218 and @xmath239 .
2 . if @xmath220 put @xmath240 and continue the loop .
3 . if @xmath222 , put @xmath241 and @xmath242 , where @xmath225 is maximal such that @xmath226 .
4 . put @xmath243 .
if @xmath228 then for @xmath244 if @xmath245 put @xmath246 .
3 . return @xmath230 .
the following algorithm computes a local projection for given @xmath179 and @xmath206 .
[ alg : localproj](localprojection ) + input : _ a finite set _ @xmath200 $ ] _ and _ @xmath201 _ _ , where @xmath202 .
_ + @xmath203 _ for @xmath179 at @xmath206 . _ 1 .
set @xmath247 , @xmath248 , @xmath249 .
2 . while @xmath76 do 1 . let @xmath250 and compute @xmath251 , @xmath252 .
if @xmath247 , @xmath253 , and an element of @xmath211 is identically zero at @xmath77 , then set @xmath254 , @xmath248 , @xmath249 and continue the loop .
3 . if @xmath247 or @xmath255 set @xmath256 else set @xmath257 .
4 . set @xmath258 .
3 . set @xmath259 .
4 . return @xmath203 .
let us first introduce an algorithm for evaluation of polynomial systems at `` partial '' sample points .
[ alg : peval](peval ) + input : _ a system _ @xmath161 _ of polynomial equations and inequalities and @xmath162 with @xmath260 . _ + 1 .
if @xmath261 or @xmath262 then return @xmath263 .
2 . if @xmath264 , where @xmath265 is one of @xmath68 or @xmath69 .
1 . if there exists a factor @xmath266 of @xmath100 such that @xmath267 $ ] and @xmath268 then return @xmath269 .
if @xmath270 $ ] return @xmath271 .
3 . return @xmath272 .
3 . if @xmath273 1 . for @xmath130 compute @xmath274 .
2 . if for some @xmath131 @xmath275 then return @xmath276 .
3 . if for all @xmath131 @xmath277 then return @xmath278 .
4 . return @xmath272 .
4 . if @xmath279 1 . for @xmath130 compute @xmath274 .
2 . if for some @xmath131 @xmath277 then return @xmath280 .
if for all @xmath131 @xmath275 then return @xmath281 .
4 . return @xmath272 .
we can now present a recursive algorithm computing cylindrical algebraic decomposition using local projections .
[ alg : lpcad](lpcad ) + input : _ a system @xmath161 _ _ of polynomial equations and inequalities and @xmath162 with @xmath163 . _ + _ a pair @xmath282 , where @xmath135 is a level-@xmath143 cylindrical subformula , _ _ , _ _ for @xmath209 , and for any cell @xmath146 if @xmath112 and for @xmath209 all elements of @xmath283 have constant signs on @xmath210 then _ 1 .
compute a disjunctive normal form @xmath284 and a conjunctive normal form @xmath285 of @xmath6 .
2 . set @xmath286 and @xmath287 .
3 . while @xmath288 do 1 . remove a tuple @xmath289 from @xmath290 .
@xmath291 are algebraic functions of @xmath292 , @xmath90 , or @xmath91 , @xmath293 , @xmath294 , @xmath295 , and the tuple represents the interval @xmath296 , 2 .
if @xmath297 set @xmath298 and set @xmath299 , where @xmath300 , else pick a rational number @xmath301 and set @xmath302 .
set @xmath303 .
3 . compute @xmath304 . if @xmath305 then set @xmath306 and @xmath307 , and go to @xmath308 .
4 . compute @xmath309 . if @xmath310 then set @xmath311 and @xmath307 , and go to @xmath308 .
5 . compute @xmath312 . for @xmath209
set @xmath313 .
compute @xmath314 .
6 . for @xmath209 set @xmath315 . 7 . if @xmath297 then set @xmath316 and go to @xmath317 .
find @xmath318and @xmath319 such that 1 .
@xmath320 and @xmath321 or @xmath322 , 2 .
@xmath323 and @xmath324 or @xmath325 , 3 .
@xmath326 and @xmath327 , 4 . either @xmath328 or @xmath329 and there are no roots of elements of @xmath211 in @xmath330
set @xmath331 .
if @xmath332 then set @xmath333 , add @xmath334 and @xmath335 to @xmath290 , and go to @xmath317 .
11 . if @xmath336 then set @xmath337 and @xmath338 .
else set @xmath339 and @xmath340 , and if @xmath341 or @xmath342 add @xmath343 to @xmath290
. 12 . if @xmath344 then set @xmath345 and @xmath346 .
else set @xmath347 and @xmath348 , and if @xmath349 or @xmath350 add @xmath351 to @xmath290
. 13 . set @xmath352
. 14 . set @xmath353 4 .
sort a by increasing values of the first element , obtaining @xmath354 . set @xmath355 .
5 . compute @xmath356 .
6 . for @xmath209 set @xmath315 . 7 .
return _ @xmath282 .
_ @xmath357 returns @xmath358 where @xmath359 is a cylindrical algebraic formula equivalent to @xmath161 .
the formula returned by algorithm [ alg : lpcad ] may involve weak inequalities , but it can be easily converted to the caf format by replacing weak inequalities with disjunctions of equations and strict inequalities . to prove correctness of algorithm [ alg : localproj ] we use the following lemmata .
[ lem : lprojmc]let _ @xmath76 _ @xmath360 _ _ , _ _ @xmath162 _ _ , _ _ a connected analytic submanifold of @xmath86 such that @xmath361 and all elements of @xmath230 are order - invariant in @xmath93 then the set @xmath362 of all elements of @xmath179 that are not identically zero on @xmath363 is analytic delineable over @xmath93 and the elements of @xmath362 are order - invariant in each @xmath362-section over @xmath93 .
suppose that @xmath364 .
step @xmath365 of algorithm [ alg : lprojmc ] guarantees that @xmath100 has a sign - invariant leading coefficient in @xmath93 .
@xmath100 does not vanish identically at any point in @xmath93 ( for @xmath219 it is ensured by step @xmath366 ; for @xmath367 it follows from irreducibility of @xmath100 ) . by theorem 3.1 of @xcite , @xmath100 is degree - invariant on @xmath93 . since @xmath368 , by theorem 2 of @xcite
, @xmath369 is analytic delineable over @xmath93 and is order - invariant in each @xmath369-section over @xmath93 .
suppose that @xmath370 and @xmath371 .
if either @xmath372 or @xmath373 has no real roots then @xmath374 is delineable on @xmath93 .
otherwise @xmath375 and hence , by theorem 2 of @xcite , @xmath374 is analytic delineable over @xmath93 .
therefore , @xmath362 is analytic delineable over @xmath93 and the elements of @xmath362 are order - invariant in each @xmath362-section over @xmath93 .
[ lem : lprojh]let _ @xmath76 _ @xmath360 _ _ , _ _ @xmath162 _ _ , _ _ a connected subset of @xmath86 such that @xmath361 and all elements of @xmath230 are sign - invariant in @xmath93 then the set @xmath362 of all elements of @xmath179 that are not identically zero on @xmath363 is delineable over @xmath93 .
suppose that @xmath376 .
let @xmath225 be maximal such that @xmath226 , and let @xmath377 .
steps @xmath365 and @xmath366 of algorithm [ alg : lprojh ] guarantee that @xmath378 in @xmath363 . by step @xmath379 and theorems 1 - 3 of @xcite , @xmath380 is delineable over @xmath93 , and hence @xmath369 is delineable over @xmath93 .
suppose that @xmath370 and @xmath371 .
if either @xmath372 or @xmath373 has no real roots then @xmath374 is delineable on @xmath93 . otherwise without loss of generality
we may assume that due to step @xmath381 @xmath230 contains all factors of @xmath382 .
by lemma 1 of @xcite and theorem 2 of @xcite , the degree of @xmath383 is constant for @xmath384 .
since @xmath100 and @xmath266 are degree - invariant in @xmath93 , by lemma 12 of @xcite , @xmath374 is delineable over @xmath93
. therefore @xmath362 is delineable over @xmath93 .
[ pro : localproj]algorithm [ alg : localproj ] terminates and returns a local projection sequence for @xmath179 at @xmath206 . to show that the algorithm terminates note that the body of the loop in step @xmath385 is executed at most @xmath386 times .
let @xmath203 be the returned sequence .
steps @xmath365 and @xmath34 ensure that @xmath204 is a finite subset of @xmath189 and @xmath205 for @xmath137 .
we will recursively construct a cell @xmath387 such that @xmath388 is the maximal connected set containing @xmath389 such that all elements of @xmath208 for @xmath209 have constant signs on @xmath390 .
moreover , for @xmath152 , the set @xmath391 of elements of @xmath211 that are not identically zero on @xmath392 is delineable over @xmath393 .
this is sufficient to prove that @xmath118 is a local projection sequence for @xmath179 at @xmath206 , because for any cell @xmath146 if @xmath207 and all elements of @xmath208 for @xmath209 have constant signs on @xmath210 then @xmath394 , by maximality of @xmath393 .
we will consider two cases depending on the value of @xmath395 when the algorithm terminated .
suppose first that when the algorithm terminated @xmath395 was @xmath15 . in this case we will additionally prove that for @xmath152 @xmath393 is an analytic submanifold of @xmath86 , all elements of @xmath204 are order - invariant in @xmath393 , and
if @xmath396 then none of the elements of @xmath211 vanishes identically at any point in @xmath393 , @xmath211 is analytic delineable on @xmath393 , and the elements of @xmath211 are order - invariant in each @xmath211-section over @xmath393 .
if @xmath397 is a root of an element of @xmath398 let @xmath399 else let @xmath400 , where @xmath88 and @xmath318 are roots of elements of @xmath398 , @xmath90 , or @xmath91 , @xmath401 , and there are no roots of @xmath398 in @xmath402 .
@xmath403 is a connected analytic submanifold of @xmath404 and all elements of @xmath398 are order - invariant in @xmath403 . since the elements of @xmath405 are irreducible , none of the elements of @xmath405 vanishes identically at any point in @xmath403 .
since all irreducible factors of elements of @xmath406 belong to @xmath398 , by lemma [ lem : lprojmc ] , @xmath405 is analytic delineable over @xmath403 and the elements of @xmath405 are order - invariant in each @xmath405-section over @xmath403 .
suppose that , for some @xmath253 , we have constructed @xmath407 satisfying the required conditions .
the conditions imply that @xmath204 is analytic delineable on @xmath407 .
let @xmath393 be the @xmath204-section or @xmath204-sector over @xmath407 which contains @xmath389 .
@xmath393 is an analytic submanifold of @xmath86 .
the elements of @xmath204 are order - invariant in @xmath393 , because they are order - invariant in each @xmath204-section over @xmath407 and nonzero in each @xmath204-sector over @xmath407 .
since all irreducible factors of elements of @xmath408 belong to @xmath409 , by lemma [ lem : lprojmc ] , @xmath391 is analytic delineable over @xmath393 and the elements of @xmath391 are order - invariant in each @xmath391-section over @xmath393 .
step @xmath410 guarantees that if @xmath396 then @xmath411 .
suppose now that when the algorithm terminated @xmath395 was @xmath17 .
let @xmath403 be as in the first part of the proof .
as before , @xmath405 is analytic delineable over @xmath403 and the elements of @xmath405 are order - invariant in each @xmath405-section over @xmath403 .
let @xmath412 be the @xmath405-section or @xmath405-sector over @xmath403 which contains @xmath413 .
@xmath412 is an analytic submanifold of @xmath414 .
the elements of @xmath405 are order - invariant in @xmath412 , because they are order - invariant in each @xmath405-section over @xmath403 and nonzero in each @xmath405-sector over @xmath403 .
since all irreducible factors of elements of @xmath415 belong to @xmath416 , by lemma [ lem : lprojmc ] , @xmath417 is analytic delineable over @xmath412 .
suppose that , for some @xmath418 , we have constructed @xmath407 satisfying the required conditions .
the conditions on @xmath407 imply that @xmath419 is delineable on @xmath407 .
let @xmath393 be the @xmath419-section or @xmath419-sector over @xmath407 which contains @xmath389 .
all elements of @xmath204 are sign - invariant in @xmath393 .
since all irreducible factors of elements of @xmath420 belong to @xmath409 , by lemma [ lem : lprojh ] , @xmath391 is delineable over @xmath393 .
since for @xmath152 , @xmath393 is the @xmath419-section or @xmath419-sector over @xmath407 which contains @xmath389 , @xmath393 is the maximal connected set containing @xmath389 such that all elements of @xmath208 for @xmath209 have constant signs on @xmath421 .
correctness and termination of algorithm [ alg : peval ] is obvious .
algorithm [ alg : lpcad ] terminates and the returned pair @xmath282 satisfies the required conditions .
let @xmath422 be the set of all polynomials that appear in @xmath6 and let @xmath423 be the hong s projection sequence @xcite for @xmath422 ( the variant of given in proposition 7 of @xcite ) .
suppose that @xmath424 and @xmath232 , where @xmath425 .
let @xmath426 .
since we assume that finite sets of polynomials are consistently ordered according to a fixed linear order in the set of all polynomials , @xmath427 for @xmath428 . hence all polynomials that appear during execution of @xmath429 are elements of @xmath430 .
in particular , @xmath88 and @xmath89 that appear in the elements of @xmath290 are roots of elements of @xmath431 , @xmath90 , or @xmath91 .
therefore , the number of possible elements of @xmath290 is finite , and hence the loop in step @xmath34 terminates .
recursive calls to @xmath432 increment @xmath78 . when @xmath249 then either step @xmath433 yields @xmath306 or step @xmath434 yields @xmath311 , and hence step @xmath435 containing the recursive call to @xmath429 is never executed .
therefore the value of @xmath78 is bounded by @xmath436 , and hence the recursion terminates .
let @xmath282 be the pair returned by @xmath429 and suppose that @xmath146 is a cell such that @xmath112 and for @xmath209 all elements of @xmath283 have constant signs on @xmath210 .
we need to show that _ _
@xmath166 let @xmath437 and @xmath438 .
we need to show that @xmath439 .
let @xmath356 , as computed in step @xmath47 .
all elements of @xmath208 have constant signs on on @xmath210 , for @xmath209 .
since none of the elements of @xmath230 vanishes identically at @xmath77 , @xmath230 is delineable over @xmath109 .
hence the @xmath230-sections and the @xmath230-sectors over @xmath109 form a partition of @xmath212 . for a tuple @xmath440 that appears on @xmath290 in any iteration of the loop in step @xmath34 put@xmath441 for each @xmath442 put @xmath443 note that each @xmath444 and @xmath445 is a union of @xmath230-sections and @xmath230-sectors over @xmath109 . put @xmath446 and @xmath447 .
we will show that in each instance of the loop in step @xmath34 @xmath448 is a partition of @xmath212 .
in the first instance of the loop in step @xmath34 @xmath449 and @xmath450 , and hence @xmath448 is a partition of @xmath212 .
we will show that this property is preserved in each instance of the loop . in each instance
a tuple @xmath440 is removed from @xmath290 and @xmath451 is added to @xmath1 . if @xmath297 in step @xmath452 then @xmath453 and the property is preserved . if @xmath332 in step @xmath454 then @xmath333 and tuples @xmath455 and @xmath456 are added to @xmath290 . since @xmath457 is a partition of @xmath444 , the property is preserved . otherwise steps @xmath458-@xmath459 are executed .
if in step @xmath458 @xmath341 or @xmath460 and @xmath342 then put @xmath461 , where @xmath462 is the tuple added to @xmath290 , else put @xmath463 .
if in step @xmath464 @xmath349 or @xmath465 and @xmath350 then put @xmath466 , where @xmath467 is the tuple added to @xmath290 , else put @xmath468 . since @xmath469 is a partition of @xmath444 , the property is preserved .
after the loop in step @xmath34 is finished @xmath290 is empty , @xmath470 , and hence @xmath471 is a partition of @xmath212 .
let @xmath442 be such that @xmath472 .
let us analyze the instance of the loop in step @xmath34 which resulted in adding @xmath473 to @xmath1 .
let @xmath474 .
suppose first that @xmath306 or @xmath311 was found in step @xmath433 or @xmath434 .
let @xmath307 , as computed in step @xmath433 or @xmath434 .
for @xmath209 , @xmath475 , and hence all elements of @xmath208 have constant signs on on @xmath476 .
therefore the set @xmath391 of elements of @xmath211 that are not identically zero on @xmath212 is delineable over @xmath109 . by definition of @xmath173 , @xmath93 is a @xmath391-section or a @xmath391-sector over @xmath109 .
hence all elements of @xmath211 have constant signs on @xmath93 . in particular , all elements of @xmath179 have constant signs on @xmath93 , and so @xmath477 .
now suppose that @xmath312 was computed in step @xmath435 .
let @xmath478 for @xmath209 , @xmath475 , and hence all elements of @xmath208 have constant signs on on @xmath476 .
as before , @xmath391 is delineable over @xmath109 , @xmath93 is a @xmath391-section or a @xmath391-sector over @xmath109 , and all elements of @xmath211 have constant signs on @xmath93 . in particular , all elements of @xmath479 have constant signs on @xmath93 . since for @xmath209 @xmath480 , all elements of @xmath481 have constant signs on on @xmath476 .
hence@xmath482 and so @xmath483 .
[ rem : zdim]the following somewhat technical improvements have been observed to improve practical performance of algorithm [ alg : lpcad ] . 1 . in step @xmath366 of algorithm [ alg : lprojmc ] in @xmath484
may be chosen arbitrarily as long as @xmath226 , hence an implementation may choose the simplest @xmath484 .
if in a recursive call to @xmath485 the initial coordinates @xmath486 correspond to single - point intervals , that is @xmath297 in step @xmath487 of the currently evaluated iteration of loop @xmath34 in all parent computations of @xmath488 for @xmath489 , then @xmath490 does not need to compute the last @xmath114 levels of projection .
instead it can return @xmath203 with @xmath491 .
computations involved in finding projections are repeated multiple times . a practical implementation needs to make extensive use of caching . in this section
we apply @xmath429 to solve the problem stated in example [ exa : mainexample ] . in step @xmath33 of @xmath492
we compute @xmath493 and @xmath494 . in the first iteration of loop @xmath34
we remove a tuple representing @xmath495 from @xmath290 and pick @xmath496 .
the calls to @xmath497 in steps @xmath433 and @xmath434 yield @xmath272 .
step @xmath435 makes a recursive call to @xmath498 . in the first iteration of loop
@xmath34 in @xmath498 we remove a tuple representing @xmath499 from @xmath290 and pick @xmath500 .
@xmath501 in step @xmath433 yields @xmath502 .
we continue on to step @xmath434 where @xmath503 yields @xmath504 .
we set @xmath311 and compute @xmath505 where @xmath506 is the set of factors of @xmath507 .
we go to step @xmath508 and set @xmath509 . in step
@xmath510 we find @xmath511 , @xmath512 , @xmath513 , and @xmath514 . in step
@xmath515 we set @xmath516 . in steps
@xmath458 and @xmath464 we add tuples representing @xmath517 and @xmath518 to @xmath290 . in step @xmath519
we obtain @xmath520 . in the second iteration of loop
@xmath34 in @xmath498 we remove a tuple representing @xmath518 from @xmath290 and pick @xmath521 .
@xmath522 in step @xmath433 yields @xmath523 .
we set @xmath306 and compute @xmath524 where @xmath506 is the set of factors of @xmath507 , @xmath525 , and @xmath526 .
we go to step @xmath508 and set @xmath509 . in step
@xmath510 we find @xmath527 , @xmath528 , and @xmath529 . in step
@xmath515 we set @xmath516 . in step
@xmath458 we add a tuple representing @xmath530 to @xmath290 . in step @xmath519
we obtain @xmath531 . in the third iteration of loop
@xmath34 in @xmath498 we remove a tuple representing @xmath530 from @xmath290 and set @xmath532 .
@xmath533 in step @xmath433 yields @xmath534 we set @xmath306 and compute @xmath524 where @xmath506 .
we go to step @xmath508 and set @xmath509 . in step
@xmath452 we set @xmath535 . in step
@xmath519 we obtain @xmath536 .
the remaining two iterations of loop @xmath34 look very similar to the last two . in step @xmath42
we obtain @xmath537 . in step @xmath47 we compute @xmath538 and in step @xmath48 we set @xmath509 .
the returned value is @xmath539 . in step @xmath435 of @xmath492
we obtain @xmath540 and @xmath541 .
@xmath542 yields @xmath543 . in step
@xmath510 we find @xmath544 , @xmath545 , @xmath546 , and @xmath547 . in steps @xmath458 and @xmath464 we add tuples representing @xmath548 and @xmath549 to @xmath290 . in step @xmath519
we obtain @xmath550 . in the second iteration of loop
@xmath34 in @xmath492 we remove a tuple representing @xmath549 from @xmath290 and pick @xmath551 . the calls to @xmath497 in steps @xmath433 and @xmath434 yield @xmath272 .
step @xmath435 makes a recursive call to @xmath552 . in the first iteration of loop
@xmath34 in @xmath552 we remove a tuple representing @xmath499 from @xmath290 and pick @xmath500 .
@xmath553 in step @xmath433 yields @xmath534 we set @xmath306 and compute @xmath554 where @xmath506 is the set of factors of @xmath35 and @xmath36 ( @xmath37 is not a part of the projection because @xmath43 and @xmath44 have no real roots ) .
we go to step @xmath508 and set @xmath509 . in step
@xmath510 we find @xmath527 and @xmath555 . in step @xmath515
@xmath230 remains empty . in step
@xmath519 we obtain @xmath556 .
the loop ends after one iteration and the returned value is @xmath557 . in step @xmath435 of @xmath492 we obtain @xmath306 and @xmath541 .
@xmath542 yields @xmath543 . in step
@xmath510 we find@xmath527 , @xmath558 , and @xmath559 . in step
@xmath458 we add a tuple representing @xmath560 to @xmath290 . in step @xmath519
we obtain @xmath561 . in the third iteration of loop
@xmath34 in @xmath492 we remove a tuple representing @xmath560 from @xmath290 and pick @xmath551 . the calls to @xmath497 in steps @xmath433 and @xmath434 yield @xmath272 .
step @xmath435 makes a recursive call to @xmath562 . in the first iteration of loop
@xmath34 in @xmath562 we remove a tuple representing @xmath499 from @xmath290 and pick @xmath500 .
@xmath563 in step @xmath433 yields @xmath564 we set @xmath306 and compute @xmath565 where , by remark [ rem : zdim ] , we can take @xmath566 .
we go to step @xmath508 and the set @xmath567 remains empty . in step
@xmath510 we find @xmath568 and @xmath569 . in step
@xmath515 we set @xmath516 . in step
@xmath454 we add tuples representing @xmath570 and @xmath571 to @xmath290 . in step @xmath519
we obtain @xmath556 . in the second iteration of loop
@xmath34 in @xmath562 we remove a tuple representing @xmath572 from @xmath290 and pick @xmath573 .
@xmath574 in step @xmath433 yields @xmath523 .
we set @xmath306 and compute @xmath575 , where , by remark [ rem : zdim ] , we can take @xmath566 .
we go to step @xmath508 and the set @xmath567 remains empty . in step
@xmath510 we find @xmath527 , @xmath576 and @xmath577 . in step @xmath515 we set @xmath516 . in step
@xmath519 we obtain @xmath578 .
the remaining iteration of loop @xmath34 look very similar to the last one . in step @xmath42
we obtain @xmath579 . in step @xmath47 we compute @xmath580 by remark [ rem : zdim ] .
the returned value is @xmath581 . in step @xmath435 of @xmath492
we obtain @xmath306 and @xmath582 .
@xmath583 yields @xmath584 . in step
@xmath452 we set @xmath585 . in step
@xmath519 we obtain @xmath586 .
the remaining two iterations of loop @xmath34 look very similar to the last two . in step @xmath42
we obtain @xmath587 and the returned value is @xmath588 .
algorithm [ alg : lpcad ] ( @xmath429 ) and the cylindrical algebraic decomposition ( @xmath589 ) algorithm have been implemented in c , as a part of the kernel of _ mathematica_. the experiments have been conducted on a linux server with a @xmath590-core @xmath591 ghz intel xeon processor and @xmath592 gb of ram available for all processes . the reported cpu time is a total from all cores used . since we do not describe the use of equational constraints in the current paper , we have selected examples that do not involve equations .
[ exa : ellipse in a square](ellipse in a square ) find conditions for ellipse @xmath596 to be contained in the square @xmath597 .
we compute a cylindrical algebraic decomposition of the solution set of@xmath598 with the free variables ordered @xmath599 .
results of experiments are given in table [ tab : benchmark - examples ] .
examples from @xcite are marked with w and the original number .
the columns marked time give the cpu time , in seconds , used by each algorithm .
the columns marked cells give the number of cells constructed by each algorithm .
the column marked wo tells whether the system is well - oriented .
experiments suggest that for systems that are not well - oriented lpcad performs better than cad . for well oriented - systems lpcad usually construct less cells than cad , but this does not necessarily translate to a faster timing , due to overhead from re - constructing projection for every cell .
however , for some of the well - oriented systems , for instance example [ exa : distance to three squares ] , lpcad is significantly faster than cad , due to its ability to exploit the boolean structure of the problem .
unfortunately we do not have a precise characterisation of such problems .
nevertheless lpcad may be useful for well - oriented problems that prove hard for the cad algorithm or may be tried in parallel with the cad algorithm . c. w. brown . constructing a single open cell in a cylindrical algebraic decomposition . in _ proceedings of the international symposium on symbolic and algebraic computation , issac 2013 _ ,
pages 133140 .
acm , 2013 .
c. chen , m. m. maza , b. xia , and l. yang .
computing cylindrical algebraic decomposition via triangular decomposition . in _ proceedings of the international symposium on symbolic and algebraic computation , issac 2009 _ , pages 95102 .
acm , 2009 .
h. hong .
an improvement of the projection operator in cylindrical algebraic decomposition . in _ proceedings of the international symposium on symbolic and algebraic computation ,
issac 1990 _ , pages 261264 .
acm , 1990 .
s. mccallum .
an improved projection for cylindrical algebraic decomposition . in b.
caviness and j. johnson , editors , _ quantifier elimination and cylindrical algebraic decomposition _ , pages 242268 .
springer verlag , 1998 .
a. strzeboski .
computation with semialgebraic sets represented by cylindrical algebraic formulas . in _ proceedings of the international symposium on symbolic and algebraic computation , issac 2010 _ , pages 6168 .
acm , 2010 .
a. strzeboski . solving polynomial systems over semialgebraic sets represented by cylindrical algebraic formulas . in _ proceedings of the international symposium on symbolic and algebraic computation ,
issac 2012 _ , pages 335342 .
acm , 2012 . | we present an algorithm which computes a cylindrical algebraic decomposition of a semialgebraic set using projection sets computed for each cell separately .
such local projection sets can be significantly smaller than the global projection set used by the cylindrical algebraic decomposition ( cad ) algorithm .
this leads to reduction in the number of cells the algorithm needs to construct .
we give an empirical comparison of our algorithm and the classical cad algorithm . |
we have analysed the evolution dynamics of a single two - level trapped ion coupled to an optical cavity interacting with a laser field and a single cavity mode . assuming that the quantum system evolves in a stochastic sequence of identical unitary transformations @xcite , the probability of generating a maximally entangled tripartite ghz state has been calculated .
we have used intraction parameters reported in a recent experiment @xcite .
as expected , the presence of decay factors in time evolution equation , causes the @xmath64 peak values to decrease with increase in the average size of the minimum time step .
the evolution dynamics of @xmath64 is also sensitive to changes in intrinsic decoherence parameter @xmath73 . besides that the energy spectrum of system hamiltonian determines the decoherence rate for a given initial state preparation . for the case at hand
, that is ion initially in its ground state occupying the lowest energy trap level , while the cavity is prepared in vacuum state , the ratio @xmath85 is crucial in determining the rate at which ghz state generation probability @xmath86 decays with time . even for a large value of @xmath87
, we obtain ghz state generation probability of @xmath88 decoherence effects due to cavity decay , heating and other random sources of decoherence on @xmath69 have not been considered . | we analyse the effects of intrinsic decoherence on the probability of generating a tripartite ghz state using a cool trapped ion coupled to a single mode of the cavity field and interacting with a resonant laser field .
milburn equation is solved for this tripartite system to obtain the time evolution of density matrix as a function of cavity - ion coupling , laser - ion coupling , and the size of unitary time step relative to the time scale determined by system parameters . starting with the system prepared initially in a separable state , density matrix
is used to calculate the probability of tripartite ghz state generation using coupling strengths reported in a recent experiment .
[ multiblock footnote omitted ] [ multiblock footnote omitted ] interest in multipartite maximally entangled states is motivated by possible use in quantum communication @xcite , quantum dense coding @xcite and quantum teleportation @xcite .
advances in experimental techniques of trapping , cooling , preparation and manipulation of ionic quantum states@xcite , through interaction with external lasers , have made the generation of maximally entangled states a possibility .
coupling of trapped ion to quantized field inside an optical cavity has also been achieved @xcite . with the ion trap placed inside a high finesse cavity , we have at hand a tripartite system with additional control mechanism offered by ion cavity coupling .
this is the basic unit of a more complex multiple function system in which quantum states of several trapped ions placed coupled to an optical cavity may be manipulated in a controlled manner . in a letter@xcite , we proposed a scheme to generate three qubit maximally entangled ghz state , using trapped ion interacting with a resonant external laser and sideband tuned single mode of a cavity field .
experimental implementation of such a proposal is constrained by the effects of quantum decoherence sources .
suppression of quantum coherence occurs due to random and unknown perturbations of system hamiltonian . in a ghz state generation experiment , the average effect of decoherence sources on probability of finding the tripartite system in ghz state is of interest . in this paper
, we investigate the effects of intrinsic decoherence on the composite system dynamics using milburn model@xcite .
milburn model of intrinsic decoherence is based on the assumption that a closed quantum system evolves in a stochastic sequence of identical unitary transformations .
the size of the unitary step along with the energy spectrum of the quantum system , determine the decoherence rate .
the model applied by milburn to free evolution of a closed quantum system , has since been used to study decoherence effects in jcm model with a single quantized field mode @xcite , decoherence in multiphoton jcm model@xcite , decoherence of open systems@xcite , decoherence of a two - level particle coupled to photon field@xcite and decoherence of quantum gate @xcite .
we solve the milburn equation exactly for a cool trapped ion coupled to a single mode of the cavity field and interacting with a resonant laser field .
the evolution dynamics of tripartite system is a function of laser - ion coupling , cavity ion coupling , and size of unitary step relative to the time scale determined by system parameters .
the systemis prepared initially in a separable state having ionic center of mass and cavity mode in respective number states , while the ion is in ground state . from the density matrix containing intrinsic decoherence effects
, we calculate the probability of tripartite ghz state generation using coupling strengths reported in recent experiments@xcite .
cold ions in a linear trap @xcite offer a promising physical system for quantum information processing , as each ion allows two qubit state manipulation . with an ion trap placed inside a high finesse optical cavity ,
we obtain a three qubit system with additional control mechanism offered by quantized cavity field .
the interaction picture hamiltonian for this system is given in a detailed form in eq .
( 4 ) of ref .
@xcite . on the time scale involving only a few cycles of unitary time evolution , the effective dimensions of three subsystems namely ionic internal state ( @xmath0 ) , state of center of mass motion ( @xmath1 ) , and cavity field state ( @xmath2 ) depend not only on the ion type but also on the relevant part of the ion - field interaction .
presently , we consider the resonant interaction of a two level cold trapped ion having transition frequency @xmath3 with an external laser of frequency @xmath4 while the cavity field is tuned to red sideband of ionic vibrational motion , @xmath5 . here
@xmath6 is the frequency of one dimensional harmonic oscillator trapping potential .
the lamb - dicke ( ld ) parameters relative to the laser field and the cavity field are denoted by @xmath7 and @xmath8 respectively . working in the lamb - dicke regime , @xmath9 and @xmath10 , the dynamics of the trapped two - level ion
is governed by interaction hamiltonian , @xmath11+\hbar g{\eta _ { c}}% \left [ \sigma _ { + } \hat{b}\hat{a}+\sigma _ { -}\hat{b}^{\dagger } \hat{a}% ^{\dagger } \right ] , \label{2.1}\ ] ] where @xmath12 and @xmath13 are creation ( annihilation ) operators for vibrational phonon and cavity field photon respectively .
the ion - phonon and ion - cavity coupling constants are @xmath14 and @xmath15 , whereas @xmath16 are the pauli operators qualifying the internal state of the ion .
to obtain time evolution of the system , we work in the basis @xmath17 @xmath18 @xmath19 and @xmath20 here @xmath21 denote the state of ionic vibrational motion and quantized cavity field , respectively . the matrix representation of interaction operator @xmath22 in the chosen basis , to the lowest order in @xmath7 and @xmath23
is @xmath24 .
\label{2.2}\ ] ] the unitary transformation that diagonalizes @xmath25 is easily obtained and yields the eigenvectors , @xmath26 ( @xmath27 ) .
milburn @xcite postulates that on sufficiently short time steps , the quantum system evolves stochastically and the change in density operator in a time interval @xmath28 is given by @xmath29 @xmath30 being some function of @xmath31 in standard quantum mechanics @xmath32 in milburn model @xmath33 , a constant minimum time step with mean frequency @xmath34 .
the evolution of density operator @xmath35 in milburn model , is governed by the equation @xmath36 , \label{2.3}\ ] ] and the rate of change of density operator to first order in @xmath37 as obtained from eq .
( [ 2.3 ] ) is @xmath38 -\frac{1}{2\hbar ^{2}\gamma } \left [ \widehat{h}% _ { i},\left [ \widehat{h}_{i},\widehat{\rho } ( t)\right ] \right ] .
\label{2.5a}\ ] ] transforming to the basis in which @xmath39 is diagonal and integrating eq .
( [ 2.5a ] ) , we get @xmath40 where @xmath41 the presence of factors of the type @xmath42 in the exponential indicates that energy spectrum of interaction operator @xmath39 plays an important role in determining the intrinsic decoherence rate . by solving eq .
( [ 2.8 ] ) , for the initial state @xmath43 , we get @xmath44 \nonumber \\ & & -\frac{(a+b)^{2}}{2}\left [ \exp \left ( -\frac{2\left ( \mu _ { mn}-a_{mn}\right ) ^{2}t}{\gamma } -i2(\mu _ { mn}-a_{mn})t\right ) \left| \phi _ { 1}\right\rangle \left\langle \phi _ { 4}\right| \right . \nonumber \\ & & \left .
+ \exp \left ( -\frac{2\left ( \mu _ { mn}-a_{mn}\right ) ^{2}t}{\gamma } % + i2(\mu _ { mn}-a_{mn})t\right ) \left| \phi _ { 4}\right\rangle \left\langle \phi _ { 1}\right| \right ] \nonumber \\ & & + \frac{(a - b)^{2}}{2}\left [ \exp \left ( -\frac{2\left ( \mu _ { mn}+a_{mn}\right ) ^{2}t}{\gamma } + i2(\mu _ { mn}+a_{mn})t\right ) \left| \phi _ { 2}\right\rangle \left\langle \phi _ { 3}\right| \right .
\nonumber \\ & & \left .
+ \exp \left ( -\frac{2\left ( \mu _ { mn}+a_{mn}\right ) ^{2}t}{\gamma } % -i2(\mu _ { mn}+a_{mn})t\right ) \left| \phi _ { 3}\right\rangle \left\langle \phi _ { 2}\right| \right ] \nonumber \\ & & \frac{(a - b)^{2}}{2}\left ( \left| \phi _ { 2}\right\rangle \left\langle \phi _ { 2}\right| + \left| \phi _ { 3}\right\rangle \left\langle \phi _ { 3}\right| \right )
\nonumber \\ & & + \frac{(a^{2}-b^{2})}{2}\left [ \exp \left ( -\frac{2\mu _ { mn}^{2}t}{\gamma } % -i2\mu _ { mn}t\right ) \left ( \left| \phi _ { 1}\right\rangle \left\langle \phi _ { 2}\right| -\left| \phi _ { 3}\right\rangle \left\langle \phi _ { 4}\right| \right ) \right .
\nonumber \\ & & + \exp \left ( -\frac{2\mu _ { mn}^{2}t}{\gamma } + i2\mu _ { mn}t\right ) \left ( \left| \phi _ { 2}\right\rangle \left\langle \phi _ { 1}\right| -\left| \phi _ { 4}\right\rangle \left\langle \phi _ { 3}\right| \right ) \nonumber \\ & & + \exp \left ( -\frac{2a_{mn}^{2}t}{\gamma } + i2a_{mn}t\right ) \left ( \left| \phi _ { 1}\right\rangle \left\langle \phi _ { 3}\right| -\left| \phi _ { 2}\right\rangle \left\langle \phi _ { 4}\right| \right ) \nonumber \\ & & \left .
+ \exp \left ( -\frac{2a_{mn}^{2}t}{\gamma } -i2a_{mn}t\right ) \left ( \left| \phi _ { 3}\right\rangle \left\langle \phi _ { 1}\right| -\left| \phi _ { 4}\right\rangle \left\langle \phi _ { 2}\right| \right ) \right ] \label{2.9}\end{aligned}\ ] ] where @xmath45 , @xmath46 , @xmath47 , and @xmath48 .
next we consider a special case with the ion in its ground state occupying the lowest energy trap state , while the cavity is prepared in vacuum state .
the density operator for the initial state is @xmath49(@xmath50 without taking into consideration the decoherence effects , the interaction time @xmath51 needed to generate the maximally entangled tripartite ghz state depends on the the ratio @xmath52 involving parameters characterizing the ion- laser and ion - cavity interaction strengths@xmath53 for interaction time @xmath54 such that @xmath55 ( @xmath56 the system is found to be in maximally entangled tripartite two mode ghz state @xcite , @xmath57 using eq .
( [ 2.9 ] ) , we can easily write down the density operator @xmath58 for @xmath59 .
next we evaluate the probability @xmath60 of finding the system in the state @xmath61
@xmath62 and obtain @xmath63 \nonumber \\ & & -\frac{1}{2}\left ( \frac{\mu _ { 11}+a_{11}}{2\mu _ { 11}}\right ) ^{2}\exp \left ( \frac{-2\left ( \mu _ { 11}-a_{11}\right ) t}{\gamma } \right ) \sin \left [ 2(\mu _ { 11}-a_{11})t\right ] . \label{2.11}\end{aligned}\ ] ] .@xmath64 peak values at @xmath65 for @xmath66.[t1 ] [ cols="^,^,^,^",options="header " , ] in the limit @xmath67 corresponding to continuous time evolution@xmath68 eq .
( [ 2.9 ] ) and eq .
( [ 2.11 ] ) reduce to decoherence free time evolution of density operator and @xmath69 respectively .
fig .
1 displays @xmath64 as a function of scaled time variable @xmath70 , for the choice @xmath71 and intrinsic decoherence parameter @xmath72 .
evidently the evolution dynamics of @xmath64 is sensitive to changes in intrinsic decoherence parameter @xmath73 . with increase in the value of @xmath73 the peaks in @xmath64 plots
are seen to get lower .
using the value @xmath74 @xcite and recalling that @xmath75 , we calculate the mean frequency @xmath34 of minimum time step .
table 1 .
lists the @xmath76 peak values at @xmath77 for the choices @xmath78 as expected , the presence of decay factors in eq .
( [ 2.11 ] ) causes the @xmath64 peak values to decrease with increase in @xmath79 for @xmath80decoherence effects dominate the scene resulting in a low @xmath64 peak value of @xmath81 .
we may recall here that for the coupling value determined by @xmath82 the angle @xmath83 corresponds to an interaction time of @xmath84 the probability of finding the system in ghz state can be measured by cavity - photon measurement combined with atomic population inversion measurement . |
transition - metal spinel compounds ab@xmath2x@xmath3 , where x is s or o , have been for many years the subject of intense experimental and theoretical activity .
the b sublattice of the spinel structure forms corner - sharing tetrahedra , giving rise to a geometrically frustrated pyrochlore lattice . in general , oxides and chalcogenides of transition metals exhibit complex charge- and spin - ordering that are often coupled with orbital - ordering .
@xcite coupled orbital - charge - spin orderings have been investigated widely for compounds of the late transition metal ions with active @xmath4 orbitals within cubic geometry , such as the cuprates and manganites . in contrast , such orderings in spinel compounds are only beginning to be studied @xcite .
spinel systems are considerably more complicated than the cuprates or manganites due to two distinct reasons .
first , the b - ions in the spinels , which are in an octahedral environment of the x anions , often possess partially - filled @xmath0 @xmath1-orbitals .
examples include v@xmath5 ions in the vanadates of zn @xcite , mn @xcite , and cd @xcite ; v@xmath6 in alv@xmath7o@xmath8 @xcite ; ti@xmath5 in mgti@xmath7o@xmath8 @xcite ; ir@xmath9 in cuir@xmath7s@xmath8 @xcite ; and rh@xmath9 in lirh@xmath7o@xmath8 @xcite .
the threefold degeneracy of the t@xmath10 orbitals , along with the weaker tendency to jahn - teller ( jt ) distortions @xcite in @xmath0-based systems lead to more complicated physics compared to @xmath4-based systems .
second , the geometric frustration in the spinel structure , mentioned above , precludes simple orderings @xcite .
very recently , an exotic orbitally - induced peierls instability has been proposed as the mechanism behind the metal - insulator transitions in the spinels like cuir@xmath7s@xmath8 @xcite .
charge - segregation of ir into formally ir@xmath5 and ir@xmath11 , accompanied by the formation of an octamer of ir@xmath11 ions with alternate short and long bonds is observed in cuir@xmath2s@xmath3 @xcite .
this unusual charge - ordering pattern has been explained within the context of an effective one - dimensionalization and peierls instability driven by orbital ordering ( oo ) .
the spinel structure of cuir@xmath7s@xmath8 consists of criss - cross chains of ir - ions with strongest overlaps between orbitals of the same kind ( i.e. , @xmath12 with @xmath12 , @xmath13 with @xmath13 ) . within the proposed mechanism ,
oo in ir@xmath9 ions leads to completely filled degenerate @xmath14 and @xmath13 orbitals and effective one - dimensional ( 1d ) quarter - filled @xmath12 bands .
the latter now undergoes a peierls instability that is accompanied by period 4 charge - ordering @xmath15 ir@xmath5/ir@xmath5/ir@xmath11/ir@xmath11 @xmath15 and long - intermediate - short - intermediate bonds , as seen experimentally@xcite .
the oo and peierls instability are thought to be coupled , as opposed to independent .
closely related phenomenologies have been proposed to explain the metal - insulators transitions in mgti@xmath7o@xmath8 @xcite and lirh@xmath7o@xmath8 @xcite .
while the proposed scenario @xcite does provide qualitative explanations , it raises many interesting questions .
for example , the bond or charge periodicities that result from peierls distortion in real quasi-1d systems depend strongly on the bandfilling . naively then
, one might be led to believe that unless the oo is complete and leads to integer occupations of the individual @xmath0 orbitals , exactly commensurate distortion periodicities , as supposedly observed in cuir@xmath7s@xmath8 @xcite and mgti@xmath7o@xmath8 @xcite , are not expected .
the exactly period 4 distortion in cuir@xmath7s@xmath8 and mgti@xmath7o@xmath8 is therefore a puzzle .
secondly , the role of electron - electron ( e - e ) interactions , which can have important consequences on these transitions , is not clear .
for example , it has been suggested that the bond distortion in mgti@xmath7o@xmath8 causes nearest - neighbor ti@xmath5 pairs to form a spin - singlet state , giving a drop in the magnetic susceptibility at low temperature@xcite .
furthermore , e - e interactions can strongly affect the distortion periodicities in 1d chains with fillings different from 1 electron per atom @xcite . addressing these issues
requires explicit calculations based on hamiltonians containing all the necessary components , that to the best of our knowledge do not exist currently .
we report here the results of explicit calculations based on a simplified model system that displays co - operative oo and peierls instability .
our model system is a two - dimensional projection of the pyrochlore lattice , a checkerboard lattice with two degenerate directional orbitals per atom .
as we discuss in the next section , the model captures the effective one - dimensionalization that occurs in the spinel lattice .
there exist also subtle differences between the our model and the theoretical picture that has been employed for the spinels @xcite , that we discuss in section [ sect_discussion ] .
we have considered two different electron densities , 1 electron per atom and @xmath16 electron per atom .
we find that a stable peierls state with 1d order occurs for an electron density of 1 electron per atom .
importantly , even for significant deviations of orbital occupancies from integer values , the peierls distortion is purely bond dimerization , as would be true in the precisely @xmath16-filled 1d band . we speculate that this is a consequence of interactions between the effective 1d chains of our model , which are not entirely independent . for the case of @xmath16 electron per atom we did not observe an orbitally - driven peierls state even for integer orbital occupancies in the absence of e - e interactions .
for the same model in the limit of infinite e - e interactions , however , a stable bond dimerization occurs .
this suggests the second main result of our paper , namely that e - e interactions are important in stabilizing orbitally - induced peierls states for the case of @xmath16-electron per atom within the orbitally degenerate checkerboard lattice .
we will discuss possible implications of this result for the spinel lattice .
the paper is divided in following subsections : in section [ sect_model ] we introduce our model hamiltonian ; in sections [ sect_1e ] and [ sect_12e ] we present numerical results for 1 and @xmath16 electron per atom respectively ; and in section [ sect_discussion ] we conclude and further discuss the relationship of our theory to spinels and related materials .
as discussed above , no quantitative calculations exist of coexisting orbital and peierls order in spinel lattices .
this is due to the huge complexity of these systems , which include three dimensionality , triply degenerate @xmath0 metal atom orbitals , possible jt distortions and frustration , in addition to e - e interactions .
we therefore construct a simpler model based on a checkerboard lattice that is easier to handle , but that is expected to bear similarities with the spinel problem .
note that although there exists a large body of literature on the consequences of frustration within the checkerboard lattice with a _
single _ atomic orbital per site @xcite , we are unaware of similar calculations on the present model which deals with doubly degenerate orbitals at each site . an important property of orbitally - driven peierls order in these materials is that the oo leads to peierls order in _ several different crystal lattice directions _ @xcite . in order to incorporate orbital degeneracy , frustration , and the possibility of peierls order in multiple directions we consider the following hamiltonian for a checkerboard lattice with doubly degenerate metal orbitals at each lattice site . @xmath17 the hamiltonian in eq .
[ spinelham ] consists of , ( i ) @xmath18 that contains the kinetic energy and the inter - ion electron - phonon ( e - p ) coupling ( eq . [ sshham ] ) , ( ii ) an oo term @xmath19 ( eq . [ ooham ] ) , and ( iii ) e - e interaction @xmath20 ( eq . [ hubbham ] ) that includes short - range e - e interactions within each site .
we describe each of these terms separately below .
@xmath18 includes electron hopping between same as well as different orbitals . in eq .
[ sshham ] , @xmath21 creates an electron of spin @xmath22 in the orbital @xmath23 of atom @xmath24 . @xmath23 and @xmath25 correspond to @xmath14 and @xmath13 orbitals , which have lobes that are oriented perpendicular to each other at each site as shown in fig .
[ lattice ] .
the inter- and intra - orbital hopping matrix elements @xmath26 are based on slater - koster parametrization of hopping integrals connecting @xmath0 orbitals @xcite . *
a * denotes a unit vector along the @xmath27 , @xmath28 , @xmath29 , or @xmath30 directions .
each bond indicated in fig .
[ lattice ] connects two orbitals at each metal ion site with two orbitals at another site , and is hence written as a 2@xmath312 matrix : @xmath32 all hopping integrals are in units of the @xmath33 matrix element ( set to -1 ) involving the two @xmath1-orbitals of the metal atoms and mediated by the @xmath34-orbital of the anion ( not shown explicitly in the figure ) in between . in the above , the small @xmath35 hoppings have been neglected .
note that electron hoppings along the @xmath27 ( @xmath28 ) direction involve only the @xmath14 ( @xmath13 ) orbitals .
thus the @xmath36-bonding among the orbitals , as opposed to the @xmath22-bonding in spinels , does not preclude the orbitally - induced effective one - dimensionalization @xcite .
the inter - ion e - p coupling in @xmath18 is written in the usual su - schrieffer - heeger ( ssh ) form @xcite . here
@xmath37 is the e - p coupling constant corresponding to the bond between atoms at @xmath24 and @xmath38 , @xmath39 the deviation of this bond from its equilibrium length , and @xmath40 the corresponding spring constant .
we include nonzero e - p couplings only along the @xmath27 and @xmath28-directions , in keeping with the peierls distortions in cuir@xmath2s@xmath3 and mgti@xmath2o@xmath3 involving only orbitals of the same kind , and assume e - p couplings of equal strength ( @xmath41 ) .
we take all spring constants @xmath42 .
the oo term @xmath19 breaks the orbital degeneracy between the two orbitals at each individual site .
@xmath43 is the number of electrons of spin @xmath22 on orbital @xmath23 of site @xmath24 , and @xmath44 .
the coordinate @xmath45 couples to the charge density difference between the two orbitals on the site , with a corresponding coupling constant @xmath46 and spring constant @xmath47 .
we fix @xmath47 to the value 1 .
as written in eq .
[ ooham ] , the relative phase of the @xmath45 at each site is unrestricted . within the model of reference
the effective one - dimensionalization is a consequence of oo alone and in principle will occur even in the absence of peierls distortion .
we have used this to determine the preferred @xmath45 mode for both 1 and @xmath16 electron per site by calculating the orbital occupations in several large lattices with open boundary condition ( obc ) . in a large obc lattice ,
the charge densities and bond orders far from the lattice edges spontaneously assume the pattern that would occur in the infinite lattice for 0@xmath48 coupling limit @xcite .
an alternate approach to determine the dominant oo mode is to calculate in a periodic lattice the energies corresponding to each mode for fixed distortion amplitude ; the dominant mode is simply the one with the lowest total energy . in the present case
we have performed both sets of calculations for both 1 and @xmath16 electron per site and have determined that the preferred oo mode in both cases is the `` checkerboard '' pattern of fig .
[ lattice](a ) , which can be parametrized in terms of a single amplitude @xmath49 with @xmath50 , where @xmath51 and @xmath52 are the @xmath27 and @xmath28 coordinates of the @xmath24th atom .
the third term in the hamiltonian , @xmath20 ( eq . [ hubbham ] ) , includes short - ranged coulomb repulsions .
@xmath54 is the on - atom coulomb repulsion for electrons in same ( different ) orbitals .
while exact diagonalization has been used successfully for many 1d , quasi-1d , and two dimensional ( 2d ) lattices involving both e - e and e - p interactions , in the present model with two orbitals per metal site the hilbert space is too large to treat any meaningful size cluster within exact diagonalization .
we will therefore consider first the non - interacting ( @xmath55 ) system , and then consider the effect of @xmath53 and @xmath56 within the unrestricted hartree - fock ( uhf ) approximation , with no assumption of the periodicity of the uhf wavefunction .
specifically , we replace the interaction terms in @xmath20 ( eq . [ hubbham ] ) by @xmath57 \end{aligned}\ ] ] @xmath58 and @xmath59 are obtained using a combination of self - consistency and simulated annealing for finding their ground state values .
the uhf approximation often gives unphysical results for large interaction strengths and we will primarily focus on small @xmath53 and @xmath56 .
we treat the oo and e - p interactions using a standard self - consistent approach derived from the equations @xmath60 the self - consistency equations derived from eq .
[ selfconsist ] are used iteratively given an initial starting distortion . in the infinite system the oo or the bond distortion would occur for infinitesimally small coupling constants @xmath46 and @xmath61 . in finite - size clusters ,
however , due to the finite - size gaps between successive energy levels , nonzero coupling constants are required before the symmetry - broken state appears . in the following we consider @xmath46 and @xmath61 close to the minimum values needed for the broken - symmetry state to occur .
we performed calculations for lattices up to 16@xmath3116 ( 256 atoms with 512 orbitals ) .
the primary difficulty in solving eq .
[ selfconsist ] is that because of the large number of quantities ( @xmath62 , @xmath39 , and uhf average charge densities ) to be determined self - consistently , the calculations often became trapped in local minima before reaching the true ground state .
in all cases we have taken care that the true ground state was reached .
below we summarize our numerical results .
these are divided into two subsections that discuss average charge densities of 1 electron per atom and @xmath16 electrons per atom , respectively .
[ sect_1e ] the oo term in our model makes the orbital occupancy of the @xmath63 and @xmath64 orbitals unequal at each site . to measure the degree of oo quantitatively ,
we calculate the majority charge density @xmath65 , defined as the charge density in the @xmath63 orbitals at @xmath66 sites ( see fig . [ lattice](a ) ) .
these orbitals form quasi-1d chains in the @xmath27 direction ( sites denoted by squares in fig .
[ dist-1e ] ) . with one electron per site , @xmath67 ranges from 0.5 to 1 , with @xmath68 indicating complete oo , and @xmath69 implying the absence of oo .
due to the symmetries present in our model , the quasi-1d chains along @xmath27 and @xmath28 directions are identical the charge density in the @xmath64 orbitals at @xmath70 sites ( denoted by circles in fig .
[ dist-1e ] ) is also @xmath67 . as an order parameter for the peierls distortion , we measure the modulation of the bond order
, @xmath71 the bond order we are interested in ( eq . [ bondorder ] ) is the expectation value of charge - transfer between orbitals of same symmetry belonging to neighboring atoms .
the charge - transfer is directly coupled to the bond distortion @xmath39 in eq .
[ sshham ] , and hence for nonzero @xmath39 the charge transfer across consecutive bonds shows periodic modulation .
the extent of modulation of @xmath72 is therefore a direct measure of the ssh distortion strength . as shown in fig .
[ dist-1e ] , we find bond order modulation along @xmath27 ( @xmath28 ) direction to involve @xmath63 ( @xmath64 ) orbitals only .
the modulation is purely period-2 ( dimerization ) with alternating strong and weak bonds , and hence we use @xmath73 , the _ difference _ between the calculated strong and weak bond orders involving orbitals of a particular symmetry , as the order parameter for the ssh distortion . because of symmetry , the amplitudes of the bond order modulations involving the @xmath63 orbitals along the @xmath27-direction , and the @xmath64 orbitals along the @xmath28-directions are identical . as discussed above ,
whether or not a co - operative orbitally - induced peierls instability occurs for @xmath74 , as well as the periodicity of the resultant bond order wave are both important issues .
we first consider the non - interacting limit ( @xmath55 ) .
the co - operative nature of the oo and bond dimerization is shown in fig .
[ half - u0 ] , where we show the results of our self - consistent calculations for 16@xmath3116 lattices . as seen in figs .
[ half - u0](a ) and (
b ) , the orbitally - induced peierls state appears only for @xmath75 , with the bond distortion a pure bond dimerization regardless of @xmath76 .
as expected for a co - operative transition , @xmath76 increases with @xmath46 , as seen in fig .
[ half - u0](a ) . for each @xmath46 there
exists an @xmath77 beyond which there occur simultaneous jumps in @xmath78 and @xmath73 ( the jump in @xmath78 becomes progressively smaller as @xmath46 increases . )
the magnitude of @xmath77 decreases with increasing @xmath46 ( see fig . [ half - u0](b ) ) to further show the cooperative effect , in fig .
[ half - u0](c ) we show the effects of ( i ) increasing @xmath61 at constant @xmath46 , and ( ii ) increasing @xmath46 at constant @xmath61 . while it is to be expected that the orbital order parameter @xmath76 increases with @xmath46 , or that the bond alternation parameter increases with @xmath61 , we find that either of the coupling constants enhances both @xmath78 and @xmath73 .
next we consider the correlated case with nonzero @xmath53 and @xmath56 . as for @xmath55 , the orbitally - driven peierls state
is again bond - dimerized . in fig .
[ half - u ] we plot @xmath78 normalized by its value in the uncorrelated system as a function of @xmath53 for several values of @xmath56 . within uhf , the combined effect of @xmath53 and @xmath56 can be to either weaken or strengthen the distortion : for fixed @xmath56 , @xmath53 tends to weaken the oo and the bond distortion , while for fixed @xmath53 , @xmath56 strengthens both order parameters . within the uhf approximation for one electron per atom ,
the effects of @xmath53 and @xmath56 cancel exactly when @xmath79 . from the hamiltonian ,
the consequence of @xmath56 is to minimize the intrasite inter - orbital coulomb repulsion , which is achieved by orbital ordering .
it is thus not surprising that @xmath56 has the same effect as @xmath46 in fig .
[ half - u ] .
the effect of @xmath53 , as seen in fig .
[ half - u ] , is however an artifact of the uhf approximation . in the case of the strictly 1d @xmath16-filled band chain with one orbital per site
, exact diagonalization and quantum monte carlo calculations have shown that the peierls bond - alternation is _ enhanced _ by the hubbard @xmath53 @xcite . in contrast , the uhf approximation predicts incorrectly that @xmath53 destroys the bond alternation in the above case @xcite .
had we been able to perform exact diagonalization in the present case , we would have found similar enhancement of the bond dimerization by @xmath53 .
this would have had a profound effect on our overall result , reducing significantly the @xmath77 or the threshold @xmath78 at which the bond dimerization appears .
the most important conclusion that follows from the above is that an orbitally - induced peierls instability can occur even for incomplete oo ( @xmath80 ) , and as long as the instability occurs at all , the bond order wave is period 2 for @xmath76 significantly less than 1 .
indeed , it is conceivable that the threshold value of @xmath78 at which the bond dimerization appears can be even smaller than 0.8 for nonzero e - e interactions .
we have found no other periodicity or evidence for soliton formation in our calculations .
an interesting aspect of the oo driven bond distortion here is the phase relationship between the bond order wave states involving the @xmath14 and @xmath13 orbitals in fig .
[ dist-1e ] .
the short bonds along the @xmath27 and @xmath28-directions occupy the same plaquettes , yielding a structure that is reminiscent of ( but different from ) the valence bond crystal obtained within the heisenberg spin - hamiltonian for the checkerboard lattice @xcite . furthermore ,
the bond dimerizations along any one direction but on different diagonals of the checkerboard lattice are strictly `` in - phase '' .
both of these indicate that while the bond dimerizations are consequences of effective one - dimensionalization , there exist strong 2d interactions in between both the criss - cross and parallel chains .
we ascribe the persistence of the bond dimerization for @xmath74 to the commensurability effect arising from the 2d interactions . in cuir@xmath2s@xmath3 and mgti@xmath7o@xmath8 , the distortion along the chain directions is not bond dimerization but a period 4 distortion @xcite .
this is as expected for a peierls transition in a 1d chain with carrier density 0.5 .
we have therefore performed self - consistent calculations within our model hamiltonian also for density 0.5 .
not surprisingly , we do obtain self - consistently an orbitally ordered state here for nonzero @xmath46 with @xmath81 . even within an essentially 100% orbitally - ordered state ( minority orbital charge density @xmath82 0.01 ) ,
however , and with very strong @xmath61 , we were unable to obtain a stable peierls - distorted state . in all cases our self - consistent simulations converged to states with disordered bond distortions and charge densities , indicating vanishing bond - charge distortion in the thermodynamic limit .
we obtained similar results after including @xmath53 and/or @xmath56 within the uhf approximation .
we conclude that for electron density away from 1 electron per atom , orbitally - driven peierls ordering does not occur within our model in the non - interacting limit or within the mean - field approach to e - e interactions .
we ascribe the absence of bond - charge distortion here to the important role played by the interaction among the crisscrossing chains within the checkerboard lattice .
we have already pointed out in section [ sect_results]a that the stabilization of the perfect period 2 distortion for the case of 1 electron per atom , even in the absence of complete oo , is a signature of such a 2d interaction .
similar 2d interactions should be relevant also for carrier density 0.5 . since within mean field theory the peierls instability in 2d is limited to carrier density of 1 , the absence of the bond - charge distortion in the present case is to be anticipated . on the other hand , we have recently shown in a series of papers that specifically for this carrier density and one orbital per site , nonzero e - e interactions can strongly stabilize bond - charge ordered states in 2d lattices @xcite . in the case of the checkerboard lattice with a single orbital per site , plaquette spin - singlet formation ( though without charge - ordering ) has similarly been found for the same carrier concentration @xcite .
it is conceivable that similar effects of e - e interactions persist in the present case with two degenerate orbitals per site .
unfortunately , performing a realistic calculation with finite @xmath53 that goes beyond the uhf approximation in the present case is beyond our computational capability .
we have therefore investigated our model hamiltonian eq .
[ spinelham ] in the limit of @xmath83 , where we assume band orbital occupancy corresponding to that for spinless fermions . fig .
[ spinless ] shows the same order parameters as in fig .
[ half - u0 ] for the spinless fermion case . while the transition occurs here for a slightly larger value of the coupling constant @xmath61 , it is otherwise identical to the transition with 1 electron per atom , viz .
, bond dimerization occurs along the diagonal directions , and oo and bond distortion reinforce each other cooperatively .
in summary , we have carried out numerical studies on the 2d checkerboard lattice with two degenerate directional orbitals per site - a model system that like the spinel compounds can in principle exhibit oo - driven peierls bond distortions and charge ordering in multiple directions .
in addition to oo and bond modulation terms , our model hamiltonian includes both intra- and inter - orbital e - e correlations that were treated within the uhf approximation .
although some of our results have strong implications for the spinels , it is useful to precisely understand the differences between the spinel and checkerboard lattices such that the applicability as well as limitations of our model can both be understood .
one difference between the two lattices is that the plaquettes in the 2d checkerboard lattice do not correspond to the tetrahedra in the spinel lattice because of the difference between horizontal and vertical bonds in fig .
[ lattice](a ) on the one hand and the diagonal bonds on the other @xcite .
what is more important in the present context is that the oo in our model is not driven by a band jt transition that destroys the degeneracies of the atomic orbitals in the model of . within our model
the two orbitals of different symmetries on a given atom are both potentially active orbitals .
we believe that our demonstration of the co - operative interaction between oo and the peierls instability in section [ sect_1e ] , where each broken symmetry enhances the other , is of direct relevance to the @xmath0-based spinel systems , where qualitative discussions have suggested similar results @xcite .
similarly , our observation that the period 2 bond distortion persists for 1 electron per atomic site even for incomplete oo , with majority charge density as low as 0.8 per orbital , may also be of significance for the spinels .
this should be particularly true for nonzero onsite hubbard interaction , which will tend to decrease the amplitude of the oo .
complete oo in the real systems cuir@xmath2s@xmath3 and mgti@xmath2o@xmath3 requires that the energy gap due to the jt distortion is significantly larger than the hubbard interaction .
it is at least equally likely that the commensurate charge and bond distortions found in the experimental systems are not due to complete oo but are consequences of the complex interactions between the crisscrossing chains in the spinel lattice , as in the checkerboard lattice .
the implication of the absence of the peierls instability for the case of @xmath16 an electron per atom in the checkerboard lattice within one - electron theory is less clear .
one possible implication is that our results for the 2d checkerboard lattice are irrelevant for the three - dimensional ( 3d ) spinel lattice because of the fundamental difference between them that has already been pointed out in the above . in cuir@xmath2s@xmath3 , the only active orbital following the oo is the @xmath12 orbital @xcite , which has been excluded within our model .
it is thus conceivable that the 1d character of the active orbitals in cuir@xmath2s@xmath3 following oo is much stronger than in the checkerboard lattice , and this is what drives the metal - insulator transition in the real system .
it is , however , equally likely that the 3d interactions between the @xmath12-based chains are as strong as the 2d interactions in the checkerboard lattice ( recall , for example , that commensurate periodicity for independent 1d chains requires complete oo , see above ) . in this case
our null result for the uncorrelated checkerboard lattice would imply non - negligible contribution of e - e interaction to the metal - insulator transitions in cuir@xmath2s@xmath3 and lirh@xmath2o@xmath3 . further theoretical work based on the 3d pyrochlore lattice as well as experimental work that determines the extent of oo in the real systems will both be necessary to completely clarify this issue . finally , assuming that e - e interactions play a role , which is subject to further investigations as discussed above , this raises an interesting question , viz .
, what ultimately is the driving force behind the metal - insulator transitions in cuir@xmath2s@xmath3 and lirh@xmath2o@xmath3 ?
three of us have argued elsewhere that for carrier concentration precisely 0.5 , there is a strong tendency to form a paired - electron crystal ( pec ) , in which there occur pairs of spin - singlet bonded sites separated by pairs of vacancies @xcite .
this tendency to spin pairing is driven by nearest neighbor antiferromagnetic ( afm ) correlations ( as would exist in a large hubbard - u system ) and is enhanced in the presence of lattice frustration .
although the original calculations are for the anisotropic triangular lattice with a single orbital per site , the same tendency to spin - singlet formation can persist also in the spinel lattice .
if the insulating state in the spinels cuir@xmath2s@xmath3 and lirh@xmath2o@xmath3 can be understood as a pec with spin - singlet pairing driven by afm correlations , it may further indicate that e - e interactions play an important role in superconductivity found in several structurally - related spinels .
whether or not e - e interactions play a role in the observed superconductivity in the spinels liti@xmath2o@xmath3 , curh@xmath2s@xmath3 , curh@xmath2se@xmath3 has remained a lingering question @xcite .
if the insulating state in this class of materials is indeed a pec , the superconducting spinels should perhaps be included among the systems in which superconductivity is driven not entirely by bcs electron - phonon coupling .
work at mississippi state university and the university of arizona was supported by the us department of energy grant de - fg02 - 06er46315 .
collaboration between the u.s .
teams and the s. n. bose national centre was supported by a travel grant from the indo - us science and technology forum grant jc/54/2007/``correlated electrons in materials '' .
the authors gratefully acknowledge discussion with d. d. sarma .
ss thanks csir for financial support . | it has been suggested that the metal - insulator transitions in a number of spinel materials with partially - filled @xmath0 @xmath1-orbitals can be explained as orbitally - driven peierls instabilities .
motivated by these suggestions , we examine theoretically the possibility of formation of such orbitally - driven states within a simplified theoretical model , a two - dimensional checkerboard lattice with two directional metal orbitals per atomic site .
we include orbital ordering and inter - atom electron - phonon interactions self - consistently within a semi - classical approximation , and onsite intra- and inter - orbital electron - electron interactions at the hartree - fock level .
we find a stable , orbitally - induced peierls bond - dimerized state for carrier concentration of one electron per atom .
the peierls bond distortion pattern continues to be period 2 bond - dimerization even when the charge density in the orbitals forming the one - dimensional band is significantly smaller than 1 .
in contrast , for carrier density of half an electron per atom the peierls instability is absent within one - electron theory as well as mean - field theory of electron - electron interactions , even for nearly complete orbital ordering .
we discuss the implications of our results in relation to complex charge , bond , and orbital - ordering found in spinels . |
the interstellar medium ( ism ) in the vicinity of the pleiades is a rich environment for the study of processes that result from the interactions between stellar photons and the gas and dust clouds of interstellar space .
the stars of this cluster were not formed out of the surrounding material visible as reflection nebulosity .
rather , the spatial association of the stars and the interstellar gas is the result of a chance encounter between the cluster and one or more approaching clouds ( white 2003 ) .
such collisions precipitate numerous radiative processes including the photoionization of atomic and molecular species , the photodissociation of molecules such as ch and h@xmath5 , and the photoelectric heating of diffuse gas by dust grains stimulated by ultraviolet radiation , all of which may help to explain some peculiarities of the ism near the pleiades .
two such peculiarities are the great strength of ch@xmath0 absorption lines observed toward many cluster members ( white 1984__a _ _ ) and the large amount of h@xmath5 in rotationally excited states ( spitzer , cochran , & hirshfeld 1974 ) .
while these anomalies may be related , a rigorous chemical model of ch@xmath0 production in the pleiades , and elsewhere in the ism , remains to be developed .
ch@xmath0 is unable to form at the low temperature of diffuse clouds because the reaction leading to its formation , c@xmath0 + h@xmath5 @xmath8 ch@xmath0 + h , is endothermic with an activation energy of @xmath9 k. elitzur and watson ( 1978 , 1980 ) showed that the above reaction can produce sufficient amounts of ch@xmath0 if the gas is heated by an interstellar shock .
however , subsequent observations failed to detect a corresponding overabundance of oh molecules resulting from a similar endothermic reaction ( federman et al .
1996__a _ _ ) .
several groups ( draine & katz 1986 ; pineau des forts et al .
1986 ) employed magnetohydrodynamic ( mhd ) shocks to lessen the problem with the overproduction of oh , but even their models yielded oh column densities in excess of observations .
the velocity shifts between ch and ch@xmath0 absorption lines predicted by these shock models have also not been detected ( e.g. , gredel , van dishoeck , & black 1993 ) .
alternate theories for a non - thermal origin of ch@xmath0 have since been proposed .
federman et al . ( 1996__b _ _ ) considered the motions of c@xmath0 ions influenced by the passage of alfvn waves against a static background of cold neutral gas .
their model predicts ch@xmath0 , ch , and oh column densities in line with observations , but requires an additional mechanism to account for observed hco@xmath0 abundances .
more recently , non - equilibrium chemistry was investigated by joulain et al .
( 1998 ) and falgarone et al .
these authors ascribe the transient heating of localized regions of the cold diffuse medium to intermittent bursts of turbulent dissipation by either mhd shocks or coherent small - scale vortices .
such models may be appropriate for many galactic sight lines , yet the observational evidence in the pleiades favors heating either by h@xmath5 dissociation or by dust photoelectron emission , rather than shocks , as the agent responsible for ch@xmath0 formation ( white 1984__b _ _ ) .
lcccccccc 23288 & 16 tau & b7 iv & 0.10 & 5.41 & 03 44 48.22 & + 24 17 22.1 & 103 @xmath10 11 & @xmath11 + 23302 & 17 tau & b6 iii & 0.05 & 3.61 & 03 44 52.54 & + 24 06 48.0 & 114 @xmath10 12 & @xmath12 + 23324 & 18 tau & b8 v & 0.05 & 5.58 & 03 45 09.74 & + 24 50 21.3 & 113 @xmath10 11 & @xmath13 + 23338 & 19 tau & b6 iv & 0.04 & 4.20 & 03 45 12.49 & + 24 28 02.2 & 114 @xmath10 14 & @xmath14 + 23408 & 20 tau & b7 iii & 0.07 & 3.81 & 03 45 49.61 & + 24 22 03.9 & 110 @xmath10 13 & @xmath12 + 23410 & & a0 v & 0.08 & 6.92 & 03 45 48.82 & + 23 08 49.7 & 103 @xmath10 11 & @xmath15 + 23432 & 21 tau & b8 v & 0.07 & 5.73 & 03 45 54.48 & + 24 33 16.2 & 119 @xmath10 13 & @xmath13 + 23441 & 22 tau & b9 v & 0.06 & 6.42 & 03 46 02.90 & + 24 31 40.4 & 109 @xmath10 11 & @xmath13 + 23480 & 23 tau & b6 iv & 0.10 & 4.11 & 03 46 19.57 & + 23 56 54.1 & 110 @xmath10 13 & @xmath14 + 23512 & & a1 v & 0.35 & 8.47 & 03 46 34.20 & + 23 37 26.5 & @xmath16 & @xmath17 + 23568 & & b9.5 v & 0.07 & 6.85 & 03 46 59.40 & + 24 31 12.4 & 150 @xmath10 23 & @xmath15 + 23629 & 24 tau & a2 v & @xmath18 & 6.30 & 03 47 21.04 & + 24 06 58.6 & @xmath16 & @xmath19 + 23630 & @xmath20 tau & b7 iii & 0.04 & 2.81 & 03 47 29.08 & + 24 06 18.5 & 113 @xmath10 13 & @xmath21 + 23753 & & b8 v & 0.04 & 5.38 & 03 48 20.82 & + 23 25 16.5 & 104 @xmath10 10 & @xmath22 + 23850 & 27 tau & b8 iii & 0.04 & 3.54 & 03 49 09.74 & + 24 03 12.3 & 117 @xmath10 14 & @xmath12 + 23862 & 28 tau & b8 vp & 0.03 & 4.97 & 03 49 11.22 & + 24 08 12.2 & 119 @xmath10 12 & @xmath23 + 23873 & & b9.5 v & 0.02 & 6.59 & 03 49 21.75 & + 24 22 51.4 & 125 @xmath10 14 & @xmath19 + 23923 & & b8 v & 0.06 & 6.12 & 03 49 43.53 & + 23 42 42.7 & 117 @xmath10 13 & @xmath13 + 23964 & & b9.5 vp & 0.11 & 6.80 & 03 49 58.06 & + 23 50 55.3 & 159 @xmath10 39 & @xmath15 + 24076 & & a2 v & 0.03 & 7.02 & 03 50 52.43 & + 23 57 41.3 & 102 @xmath10 10 & @xmath15 + tau [ @xmath24 , @xmath25 .
the label next to each symbol is either the last three digits of the hd number or the flamsteed or bayer designation . decreasing symbol size
denotes three magnitude ranges : @xmath26 , @xmath27 , and @xmath28 .
this key is used in figures 6 , 7 , and 8.,scaledwidth=40.0% ] precise measurements of cn , ch@xmath0 , and ch column densities in the pleiades interstellar gas can offer new insight into the chemical reaction networks active in these diffuse clouds . at the same time , highly sensitive observations of and , to a lesser extent , can be used to trace the kinematic effects of the interaction occurring between interstellar gas and the stellar radiation field in great detail and on large spatial scales .
the earlier investigation by white ( 1984__a _ _ ) , comprising spectra for 15 pleiades members , analyzed interstellar absorption from the above atomic and molecular species , but the moderate velocity resolutions ( @xmath2 3@xmath298 km s@xmath3 ) allowed detections of only one component per sight line .
later studies capable of detecting many line - of - sight components , such as the ultra - high resolution survey by crane et al .
( 1995 ) of interstellar ch@xmath0 and ch and the high - resolution survey by welty et al .
( 1996 ) of k , included only a few of the brightest pleiades members in their sample .
most recently , high - resolution observations were made by white et al .
( 2001 ) of 36 pleiades stars , both members and nonmembers , in the d lines and of 12 of these stars in the ultraviolet doublet .
the full analysis of these data ( white 2003 ) revealed considerable complexity and star - to - star variation in the atomic gas traced by neutral sodium , leading to an extensive spatial schematic of cloud - cluster interactions in which two clouds were presumed to be interacting with the uv radiation field of the pleiades and also with each other .
one of the primary motivations for the present investigation was to obtain equally high quality data for the other important optical tracers of the ism toward a large number of targets in the pleiades so that a complete picture of the interaction between interstellar gas clouds and the stars of the cluster may be constructed . in this paper , high - resolution , high signal to noise observations of cn , , , ch@xmath0 , and
ch toward 20 pleiades members allow us to study cloud - cluster interactions and the implications for the chemistry of the local ism with a precision unmatched by previous investigations .
we describe our observations and detail the process of data reduction in 2 . in 3 , we present the results of gaussian fitting the observed profiles and compare our measurements to those from the literature .
the analysis of our observations appears in 4 , with particular attention paid to the spatial distribution of atomic and molecular velocity components ( 4.1 ) .
we examine / column density ratios in 4.2 and derive physical conditions in 4.3 . in 5 , we interpret our findings in light of the conceptual schematic of cloud - cluster interactions offered by white ( 2003 ) and summarize our principal results in 6 .
we observed 20 stars in the pleiades using the high - resolution mode ( cs21 ) of the 2dcoud spectrograph ( tull et al .
1995 ) on the harlan j. smith 2.7 m telescope at mcdonald observatory in january , 2002 .
the choice of sight lines came from the list in white et al .
( 2001 ) , from which only those stars with secure membership in the pleiades cluster were included in our survey .
stellar and observational data for the stars in our sample appear in table 1 . specifically , the hd number , name , spectral type , reddening , @xmath30 magnitude , equatorial coordinates ( j2000 ) , distance , and total exposure time on each star are presented .
the spectral types and @xmath31(@xmath32 ) values come from white et al .
( 2001 ) , while the @xmath30 magnitudes and equatorial coordinates were obtained from the simbad database , operated at cds , strasbourg , france .
it is most appropriate to list @xmath30 magnitudes because our instrumental setup had a central wavelength near 4000 .
the parallax for each star was obtained from _ hipparcos _ measurements ( perryman 1997 ) , when available , and used to calculate the distance given in table 1 .
additional information on the program stars is available from table 2 of white et al .
for reference , figure 1 shows the distribution of observed sight lines in the pleiades centered on @xmath20 tau .
the cross - dispersed echelle spectrometer of the 2dcoud spectrograph was configured in a manner identical to that described by pan et al .
( 2004 ) for their mcdonald observatory spectra . when combined with a 2048 @xmath33 2048 ccd , this configuration made possible the search for interstellar absorption from many species with a single exposure .
in particular , the spectral coverage allowed for the detection of absorption features from cn near 3874 , k at 3933 , @xmath344226 , ch@xmath35 @xmath344232 , and ch @xmath344300 .
each star was observed at least twice with a maximum exposure time of 30@xmath36 per frame to minimize the effect of cosmic rays .
calibration exposures for dark current were obtained the first night of the run , while exposures for bias - correction and flat - fielding were taken each night and th - ar comparison spectra were obtained every 2 to 3 hours . by measuring the width of thorium emission lines in the comparison spectra , the resolution for these observations is determined to be about 1.7 km s@xmath3 , equivalent to a resolving power of @xmath37 .
standard iraf routines were used for bias - correction , cosmic - ray removal , and flat - fielding of the raw image data .
one - dimensional spectra were extracted from the processed images and were calibrated in wavelength using the th - ar comparison spectra and then doppler - corrected . in regions where interstellar absorption lines were expected ,
individual spectra were cut from each stellar exposure with sufficient continua on both sides , typically 2 .
all spectra for the same species toward the same star were summed to yield a higher signal - to - noise ratio ( snr ) in the final spectrum .
the snrs obtained in this way yielded 3-@xmath38 upper limits of about 2 m on cn absorption and about 1 m above 4000 .
further details concerning the many steps in the data reduction process can be found by consulting pan et al .
( 2004 ) , with the caveat that no smoothing or additional processing was performed on the data presented here .
stellar spectra were normalized to unity by fitting legendre polynomials , of the lowest possible order , to regions free of interstellar absorption .
this process proceded without complication except in the case of hd 23964 .
the spectra for all of our targeted species toward this star showed strong stellar features near the position of the expected interstellar lines . for k
, we were able to normalize the stellar feature because of the strength of the interstellar lines and their location away from the core of the stellar profile .
but for cn , , ch@xmath35 , and ch , the stellar features could not be removed by normalization .
figures 2@xmath294 present the final normalized spectra for k , ch@xmath35 , and ch , respectively .
the panels are arranged to approximate the spatial relationship among sight lines so that a comparison might be made of absorption profiles in different regions of the cluster ( see fig .
2 ) . figure 5 presents the final normalized spectrum for cn absorption toward hd 23512 , the only sight line in our sample with a detectable amount of this molecule . because , no absorption ( stronger than @xmath2 1 m ) is detected along any of the observed sight lines , the spectra for this species are not given .
these negative results are not unexpected considering the survey of by welty , hobbs , & morton ( 2003 ) which found predominantly quite weak lines .
the resolution and signal to noise of our observations were high enough to allow the detection of multiple components of k and ch@xmath0 and to permit the first positive detections of cn and ch toward a few of the observed stars .
the lines were relatively weak , enabling us to fit simple gaussians to the components , yielding values for the equivalent width ( @xmath39 ) , radial velocity in the local standard of rest ( @xmath4 ) , and full width at half - maximum ( fwhm ) for each component .
the doppler parameter ( @xmath1-value ) was obtained from the measured line width ( fwhm ) , corrected for the instrumental width determined from thorium emission lines in the th - ar spectra .
uncertainties in @xmath39 were derived from the rms variations in the normalized continuum and the width of the line for a specific component .
they do not include errors in the placement of the continuum , but this is assumed to have a negligible effect .
when no component was detected , 3-@xmath38 upper limits were calculated by adopting a typical width for the species in question .
subsequent profile synthesis including @xmath40-doubling for the ch features yielded no significant change in fitted @xmath39 .
thus , the analysis of this paper rests solely on the results of our gaussian fits .
the moderately complex component structures of the k and ch@xmath0 absorption profiles were constrained by the derived @xmath1-values .
in general , no component narrower than 0.4 km s@xmath3 or broader than 2.5 km s@xmath3 was accepted , where the upper limit to @xmath1 is 1.5 times the mean value for found by pan et al .
we note that this restriction plays at most a minor role in the component analysis . in some cases
, it was necessary to compare the velocity of a ch@xmath0 component to that of k to determine its legitimacy . as discussed in some detail in pan et al .
( 2004 ) , because is the most widely distributed of the observed species , if any other species has a component at a particular @xmath41 , then should also have a component at that velocity . based on our resolution ,
an agreement within 1.5 km s@xmath3 was deemed adequate .
the final component structure for a given line of sight resulted from using the fewest number of components , subject to the above criteria , which adequately fit the observed profile and left the residuals indistinguishable from the noise in the continuum .
naturally , this approach presumes that the velocity structure along a line of sight is constant across species . to test this assumption ,
we computed the difference between the velocities found for associated ch@xmath0 and components , @xmath42(ch@xmath43 ( ) . the average difference in velocity for components detected in both species is + 0.3 km s@xmath3 . to the extent that this result is comparable to the estimated uncertainty ( @xmath44 [email protected] km s@xmath3 )
, we can be reasonably assured that our fits reveal the true component structure for a given line of sight .
km s@xmath3 and @xmath45 km s@xmath3 on this scale , respectively .
the figure shows positive detections of both the r(0 ) and r(1 ) lines .
the p(1 ) line is below the detection limit.,scaledwidth=40.0% ] cccccc + 1 & 10.7 @xmath10 0.7 & @xmath46 0.9 & 16.0 @xmath10 0.4 & 0.5 @xmath10 0.3 & @xmath46 1.8 + 2 & 10.8 @xmath10 1.4 & @xmath46 2.4 & 19.5 @xmath10 1.1 & @xmath46 4.4 & @xmath46 5.5 + 3 & 15 @xmath10 3 & & 20 @xmath10 2 & @xmath46 13 & + 4 & & & 18.5 & & + + 1 & 6.5 @xmath10 0.5 & @xmath46 0.7 & 1.9 @xmath10 0.3 & @xmath46 0.6 & @xmath46 1.4 + 2 & 6.0 @xmath10 1.5 & @xmath46 3.5 & @xmath46 2.4 & @xmath46 2.0 & @xmath46 3.6 + 5 & 4.4 @xmath10 0.4 & & & & + 6 & 4 & & & & + + 1 & 11.5 @xmath10 0.5 & @xmath46 0.8 & 3.0 @xmath10 0.3 & @xmath46 0.8 & @xmath46 1.2 + 2 & @xmath46 13 & @xmath46 4.0 & @xmath46 4.6 & @xmath46 4.0 & @xmath46 7.7 + + 1 & 11.6 @xmath10 0.5 & @xmath46 0.6 & 2.8 @xmath10 0.2 & @xmath46 0.5 & @xmath46 1.2 + 2 & 12.3 @xmath10 1.5 & @xmath46 4.5 & 4.0 @xmath10 1.3 & @xmath46 4.1 & @xmath46 3.7 + 3 & @xmath46 13 & & @xmath46 13 & @xmath46 13 & + 6 & 9 & & & & + + 1 & 9.0 @xmath10 0.4 & @xmath46 0.8 & 24.4 @xmath10 0.3 & @xmath46 0.7 & @xmath46 1.5 + 2 & 12.4 @xmath10 1.1 & @xmath46 2.2 & 22.2 @xmath10 0.5 & @xmath46 1.9 & @xmath46 2.1 + 4 & & & 20.4 & & + 6 & 8 & & & & + 7 & & & 24.2 & @xmath46 0.2 & + 8 & & & 23.3 @xmath10 0.3 & & + 9 & & & 20.2 @xmath10 1.8 & 1.3 @xmath10 0.5 &
+ 10 & & & 23.5 & & + 11 & & & 31.0 & & + 12 & & & 25.4 & &
+ + 1 & 12.4 @xmath10 0.9 & @xmath46 1.2 & 18.2 @xmath10 0.7 & @xmath46 1.1 & @xmath46 2.4 + 2 & 13.4 @xmath10 1.8 & @xmath46 6.9 & 20.0 @xmath10 1.5 & @xmath46 7.9 & @xmath46 5.7 + 3 & 15 @xmath10 2 & & 25 @xmath10 3 & @xmath46 13 &
+ + 1 & 7.3 @xmath10 0.6 & @xmath46 1.0 & 14.0 @xmath10 0.4 & @xmath46 0.9 & @xmath46 1.6 + 2 & 17.7 @xmath10 3.2 & @xmath46 8.1 & @xmath46 8.7 & @xmath46 6.7 & @xmath46 10 + + 1 & 9.0 @xmath10 0.4 & @xmath46 0.6 & 15.0 @xmath10 0.4 & @xmath46 0.6 & @xmath46 1.3 + 2 & 10.3 @xmath10 1.1 & @xmath46 2.9 & 14.1 @xmath10 0.5 & @xmath46 2.6 & @xmath46 3.5 + 3 & @xmath46 13 & & 25 @xmath10 5 & @xmath46 13 & + 4 & & & 14.2 & & + 6 & 7 & & & & + 7 & & & 16.2 & @xmath46 3.1 & + 8 & & & 14.0 @xmath10 0.2 & & + 9 & & & 12.5 @xmath10 2.2 & @xmath46 2.9 & + 11 & & & 16.0 & & + + 1 & 11.2 @xmath10 1.7 & @xmath46 2.2 & 5.1 @xmath10 0.6 & 4.3 @xmath10 0.6 & 8.7 @xmath10 0.7 + 2 & @xmath46 22 & @xmath46 13 & @xmath46 14 & 23 @xmath10 5 & @xmath46 29 + 13 & & & & 5.3 @xmath10 0.8 & @xmath46 10.0 + + 1 & 12.5 @xmath10 1.0 & @xmath46 1.6 & 18.4 @xmath10 0.6 & 1.3 @xmath10 0.4 & @xmath46 3.3 + 2 & 15.6 @xmath10 3.0 & @xmath46 7.9 & 15.5 @xmath10 2.6 & @xmath46 7.7 & @xmath46 8.7 + + 1 & 12.5 @xmath10 0.4 & @xmath46 0.4 & 2.0 @xmath10 0.1 & 0.4 @xmath10 0.1 & @xmath46 0.9 + 2 & 15.0 @xmath10 1.1 & @xmath46 3.3 & 3.8 @xmath10 0.7 & @xmath46 2.8 & @xmath46 3.7 + 5 & 12.1 @xmath10 0.7 & & & & + 5 & 14.4 @xmath10 1.2 & &
& & + 6 & 11 & & & & + 7 & & & 1.7 & & + 9 & & & 1.6 @xmath10 0.4 & @xmath46 1.8 & + 14 & 13.6 @xmath10 0.2 & & & & + 15 & 8 & @xmath46 0.64 & & &
+ + 1 & 11.0 @xmath10 0.6 & @xmath46 0.9 & 3.5 @xmath10 0.2 & @xmath46 0.7 & @xmath46 1.2 + 2 & 10.6 @xmath10 1.0 & @xmath46 4.2 & 4.0 @xmath10 0.9 & @xmath46 4.5 & @xmath46 5.5 + 4 & & & 7.6 & & + + 1 & 17.7 @xmath10 0.4 & @xmath46 0.7 & 4.1 @xmath10 0.3 & 0.4 @xmath10 0.2 & @xmath46 1.1 + 2 & 12.1 @xmath10 1.2 & @xmath46 3.9 & 5.0 @xmath10 0.7 & @xmath46 3.1 & @xmath46 4.1 + 3 & 13 @xmath10 1 & & @xmath46 13 & @xmath46 13 & + 6 & 12 & & & & + 7 & & & 4.2 & & + + 1 & 17.0 @xmath10 0.3 & @xmath46 0.5 & 4.1 @xmath10 0.2 & 0.6 @xmath10 0.1 & @xmath46 0.8 + 2 & 13.8 @xmath10 1.6 & @xmath46 4.5 & @xmath46 2.9 & @xmath46 3.5 & @xmath46 5.1 + cccccc
+ 1 & 12.7 @xmath10 0.7 & @xmath46 1.0 & 2.4 @xmath10 0.3 & @xmath46 0.8 & @xmath46 1.4 + 2 & 16.9 @xmath10 1.5 & @xmath46 5.8 & @xmath46 11 & @xmath46 5.1 & @xmath46 11 + as seen in figure 2 , interstellar absorption from k is detected along all twenty sight lines , resulting in 65 individual components .
we note that our study detects 50 percent more absorption components in than white et al .
( 2001 ) found in their observations for the same directions .
figure 3 shows that ch@xmath0 absorption is just as pervasive , yet only 23 components are identified .
no detection of ch@xmath0 toward hd 23964 is possible because the core of the stellar line discussed in 2 directly coincides with the expected position of the interstellar feature .
marginal detections of ch toward five of the observed stars are indicated in figure 4 .
a stronger component is detected toward hd 23512 , a sight line which passes through a known molecular cloud ( federman & willson 1984 ) .
the spectrum of cn toward this star ( figure 5 ) clearly shows the absorption features of the r(0 ) and r(1 ) transitions of this molecule , but the weaker p(1 ) line is below the detection limit .
lccc ca ii & 3933.663 & 0.6346 & 1 + ch@xmath35 & 4232.548 & 0.00545 & 2 + ch & 4300.313 & 0.00510 & 3 + cn & 3874.00 & 0.0228 & 3 + & 3874.61 & 0.0342 & 3 + & 3875.76 & 0.0114 & 3 + cccccccccccccccc + 1.8 @xmath10 0.5 & @xmath47 & 1.9 & & @xmath46 1.8 & & & & @xmath46 1.2 & & & & @xmath46 0.9 & & & @xmath46 5.0 + 4.6 @xmath10 0.5 & + 5.1 & 1.4 & & @xmath46 1.8 & & & & 0.6 @xmath10 0.4 & + 6.6 & 1.6 & & 2.8 @xmath10 0.3 & + 6.4 & 1.3 & 6.1 @xmath10 0.9 + 6.4 @xmath10 0.5 & + 8.5 & 2.0 & & 21.7 @xmath10 0.6 & + 9.1 & 1.9 & & @xmath46 1.2 & & & & 2.5 @xmath10 0.3 & + 9.3 & 1.1 & 3.9 @xmath10 0.6 + + 0.8 @xmath10 0.2 & + 3.3 & 0.4 & & @xmath46 0.6 & & & & @xmath46 0.7 & & & & @xmath46 0.2 & & & @xmath46 2.5 + 4.3 @xmath10 0.2 & + 6.7 & 1.4 & & 0.9 @xmath10 0.2 & + 7.5 & 1.6 & & @xmath46 0.7 & & & & 2.0 @xmath10 0.1 & + 7.0 & 1.1 & 4.6 @xmath10 0.3 + 1.8 @xmath10 0.4 & + 9.7 & 1.5 & & 1.3 @xmath10 0.2 & + 10.3 & 1.6 & & @xmath46 0.7 & & & & 0.4 @xmath10 0.1 & + 9.4 & 0.7 & 2.2@xmath10 0.7 + 0.8 @xmath10 0.2 & + 14.0 & 1.3 & & @xmath46 0.6 & & & & @xmath46 0.7 & & & & @xmath46 0.2 & & & @xmath46 2.5 + + 1.9 @xmath10 0.4 & + 4.3 & 1.6 & & @xmath46 1.2 & & & & @xmath46 1.0 & & & & @xmath46 0.4 & & & @xmath46 2.1 + 9.8 @xmath10 0.4 & + 6.7 & 1.6 & & @xmath46 1.2 & & & & @xmath46 1.0 & & & & 3.6 @xmath10 0.5 & + 7.4 & 1.0 & 3.7 @xmath10 0.5 + 2.3 @xmath10 0.4 & + 9.8 & 1.6 & & 3.6 @xmath10 0.4 & + 9.6 & 1.5 & & @xmath46 1.0 & & & & 0.8 @xmath10 0.1 & + 10.9 & 0.9 & 3.5 @xmath10 0.7 + + 1.5 @xmath10 0.4 & @xmath48 & 1.8 & & @xmath46 0.6 & & & & @xmath46 0.6 & & & & @xmath46 3.0 & & & @xmath46 20.0 + 3.4 @xmath10 0.4 & + 5.7 & 1.6 & & 3.3 @xmath10 0.2 & + 6.9 & 1.8 & & @xmath46 0.6 & & & & 4.0 @xmath10 1.0 & + 6.7 & 1.0 & 11.8 @xmath10 3.2 + 9.2 @xmath10 0.4 & + 7.6 & 1.9 & & @xmath46 0.6 & & & & @xmath46 0.6 & & & & 6.0 @xmath10 1.1 & + 8.3 & 0.8 & 6.5 @xmath10 1.2 + + 9.3 @xmath10 0.4 & + 6.6 & 1.7 & & 15.7 @xmath10 0.3 & + 7.2 & 1.5 & & @xmath46 0.8 & & & & 8.0 @xmath10 0.4 & + 6.9 & 1.0 & 8.6 @xmath10 0.6 + 1.8 @xmath10 0.2 & + 8.5 & 0.9 & & 16.0 @xmath10 0.3 & + 9.0 & 1.6 & & @xmath46 0.8 & & & & 2.4 @xmath10 0.2 & + 8.9 & 0.9 & 13.3 @xmath10 1.8 + + 2.8 @xmath10 0.4 & @xmath49 & 0.9 & & @xmath46 1.2 & & & & @xmath46 1.2 & & & & 0.6 @xmath10 0.2 & @xmath295.2 & 0.2 & 2.1 @xmath10 0.8 + 7.7 @xmath10 0.5 & + 6.9 & 1.4 & & 2.7 @xmath10 0.4 & + 6.9 & 2.1 & & @xmath46 1.2 & & & & 12 .
@xmath10 0.8 & + 7.1 & 1.1 & 15.6 @xmath10 1.4 + + 3.8 @xmath10 0.6 & @xmath50 & 1.6 & & @xmath46 2.1 & & & & @xmath46 1.3 & & & & @xmath46 1.9 & & & @xmath46 5.0 + 6.8 @xmath10 0.8 & + 5.9 & 2.2 & & 4.9 @xmath10 0.6 & + 7.0 & 2.3 & & @xmath46 1.3 & &
& & 4.1 @xmath10 0.7 & + 6.9 & 1.0 & 6.0 @xmath10 1.2 + 4.3 @xmath10 0.6 & + 8.8 & 1.8 & & 18.7 @xmath10 0.8 & + 9.4 & 2.1 & & @xmath46 1.3 & & & & 1.3 @xmath10 0.6 & + 9.2 & 1.1 & 3.0 @xmath10 1.4 + + 1.3 @xmath10 0.5 & + 3.5 & 1.4 & & @xmath46 1.8 & & & & @xmath46 1.1 & & & & @xmath46 1.4 & & & @xmath46 10.8 + 5.9 @xmath10 0.5 & + 6.4 & 1.4 & & @xmath46 1.8 & & & & @xmath46 1.1 & & & & 4.0 @xmath10 0.5 & + 6.8 & 1.0 & 6.8 @xmath10 1.0 + 1.5 @xmath10 0.4 & + 9.2 & 1.2 & & 18.6 @xmath10 0.6 & + 8.6 & 2.0 & & @xmath46 1.1 & & & & 0.9 @xmath10 0.4 & + 9.2 & 1.0 & 6.0 @xmath10 3.1 + + 1.0 @xmath10 0.2 & @xmath51 & 1.9 & & @xmath46 1.2 & & & & @xmath46 0.7 & & & & @xmath46 0.4 & & & @xmath46 4.0 + 0.7 @xmath10 0.2 & + 3.1 & 1.5 & & @xmath46 1.2 & & & & @xmath46 0.7 & & & & @xmath46 0.4 & & & @xmath46 5.7 + 4.2 @xmath10 0.2 & + 6.8 & 1.7 & & 1.3 @xmath10 0.4 & + 7.5 & 2.3 & & @xmath46 0.7 & & & & 3.0 @xmath10 0.5 & + 7.5 & 1.0 , 3.4 & 7.1 @xmath10 1.2 + 3.8 @xmath10 0.2 & + 9.4 & 1.6 & & 18.4 @xmath10 0.4 & + 10.2 & 2.2 & & @xmath46 0.7 & & & & 1.0 @xmath10 0.1 & + 10.6 & 0.6 & 2.6 @xmath10 0.3 + 0.9 @xmath10 0.2 & + 13.7 & 1.5 & & @xmath46 1.2 & & & & @xmath46 0.7 & & & & @xmath46 0.4 & & & @xmath46 4.4 + + 5.8 @xmath10 1.5 & + 6.2 & 1.5 & & @xmath46 2.4 & & & & @xmath46 2.4 & & & & 5.9 @xmath10 3.1 & + 7.4 & 0.5 & 10.2 @xmath10 6.0 + 7.8 @xmath10 1.6 & + 9.2 & 1.6 & & 6.2 @xmath10 0.8 & + 9.1 & 1.7 & & 5.4 @xmath10 0.8 & + 9.8 & 1.6 & & @xmath52 60 .
& + 10.1 & 0.6 & @xmath52 76.9 + + 6.3 @xmath10 0.6 & + 5.5 & 1.6 & & @xmath46 3.0 & & & & @xmath46 1.5 & & & & @xmath46 2.5 & & & @xmath46 4.0 + 6.3 @xmath10 0.6 & + 7.6 & 1.8 & & @xmath46 3.0 & & & & 1.6 @xmath10 0.5 & + 7.5 & 1.6 & & 9.4 @xmath10 1.0 & + 7.1 & 1.0 & 14.9 @xmath10 2.1 + 2.5 @xmath10 0.8 & + 10.0 & 2.5 & & 25.6 @xmath10 1.0 & + 8.8 & 2.1 & & @xmath46 1.5 & & & & 1.8 @xmath10 0.6 & + 9.5 & 0.6 & 7.2 @xmath10 3.3 + + 1.2 @xmath10 0.2 & @xmath53 & & & @xmath46 0.6 & & & & @xmath46 0.8 & & & & @xmath46 2.0 & & & @xmath46 16.7 + 3.1 @xmath10 0.5 & + 4.3 & 1.4 & & @xmath46 0.6 & & & & @xmath46 0.8 & & & & @xmath46 2.0 & & & @xmath46 6.4 + 11.1 @xmath10 0.5 & + 7.2 & 1.5 & & 2.5 @xmath10 0.2 & + 7.6 & 1.0 & & @xmath46 0.8 & & & & 9.7 @xmath10 0.9 & + 7.1 & 2.6 , 0.9 & 8.7 @xmath10 0.9 + 1.5 @xmath10 0.4 & + 11.6 & 1.1 & & @xmath46 0.6 & & & & @xmath46 0.8 & & & & @xmath46 2.0 & & & @xmath46 13.3 + + 2.0 @xmath10 0.2 & @xmath54 & 1.2 & & @xmath46 0.3 & & & & @xmath46 0.3 & & & & 0.1 @xmath10 0.02 & @xmath293.4 & 1.9 & 0.5 @xmath10 0.1 + 2.3 @xmath10 0.2 & + 4.8 & 1.3 & & @xmath46 0.3 & & & & @xmath46 0.3 & & & & 0.4 @xmath10 0.05 & + 5.6 & 3.8 & 1.7 @xmath10 0.3 + 9.7 @xmath10 0.3 & + 7.2 & 1.3 & & 2.4 @xmath10 0.1 & + 7.2 & 1.4 & & 0.5 @xmath10 0.1 & + 6.0 & 1.5 & & 7.5 @xmath10 0.2 & + 7.5 & 1.1 & 7.7 @xmath10 0.3 + 1.2 @xmath10 0.2 & + 11.6 & 1.3 & & @xmath46 0.3 & & & & @xmath46 0.3 & & & & @xmath46 0.1 & & & @xmath46 0.8 + + 2.5 @xmath10 0.5 & + 3.7 & 1.7 & & @xmath46 0.6 & & & & @xmath46 0.8 & & & & 1.5 @xmath10 0.2 & + 4.0 & 0.3 & 6.0 @xmath10 1.4 + 9.2 @xmath10 0.4 & + 7.3 & 1.2 & & 4.2 @xmath10 0.2 & + 8.1 & 1.7 & & @xmath46 0.8 & & & & 7.2 @xmath10 0.3 & + 7.6 & 1.2 & 7.8 @xmath10 0.5 + 1.8 @xmath10 0.5 & + 10.6 & 2.2 & & @xmath46 0.6 & & & & @xmath46 0.8 & & & & @xmath46 0.5 & & & @xmath46 2.8 + cccccccccccccccc + 2.2 @xmath10 0.2 & @xmath55 & 0.8 & & @xmath46 1.2 &
& & & @xmath46 0.6 & & & & 0.1 @xmath10 0.01 & @xmath294.1 & 1.6 & 0.4 @xmath10 0.1 + 0.9 @xmath10 0.2 & + 4.3 & 1.4 & & @xmath46 1.2 & & & & @xmath46 0.6 & & & & @xmath46 0.03 & & & @xmath46 0.3 + 17.7 @xmath10 0.3 & + 7.1 & 1.3 & & 4.9 @xmath10 0.4 & + 7.2 & 2.1 & & 0.5 @xmath10 0.2 & + 7.2 & 1.6 & & 20 .
@xmath10 1.1 & + 7.2 & 3.8 , 0.9 & 11.3 @xmath10 0.6 + 1.9 @xmath10 0.2 & + 11.0 & 1.8 & & @xmath46 1.2 & & & & @xmath46 0.6 & & & & 0.7 @xmath10 0.2 & + 9.0 & 0.6 & 3.7 @xmath10 1.1 + + 2.0 @xmath10 0.1 & @xmath56 & 1.0 & & @xmath46 0.6 & & & & @xmath46 0.3 & & & & @xmath46 0.6 & & & @xmath46 3.0 + 0.7 @xmath10 0.1 & + 4.0 & 1.2 & & @xmath46 0.6 & & & & @xmath46 0.3 & & & & @xmath46 0.6 & & & @xmath46 8.6 + 17.9 @xmath10 0.3 & + 7.3 & 1.4 & & 4.9 @xmath10 0.2 & + 7.4 & 2.0 & & 0.7 @xmath10 0.1 & + 7.4 & 1.6 & & 4.7 @xmath10 0.2 & + 6.8 & 1.0 & 2.6 @xmath10 0.1 + 1.4 @xmath10 0.2 & + 11.5 & 2.1 & & @xmath46 0.6 & & & & @xmath46 0.3 & & & & @xmath46 0.6 & & & @xmath46 4.3 + + 16.2 @xmath10 1.0 & + 5.8 & 1.8 & & @xmath46 2.1 & & & & @xmath46 1.2 &
& & & 7.9 @xmath10 0.3 & + 6.6 & 1.5 & 4.9 @xmath10 0.4 + 5.0 @xmath10 0.7 & + 8.1 & 1.6 & & 4.4 @xmath10 0.7 & + 7.7 & 2.1 & & @xmath46 1.2 & & & & @xmath46 1.0 & & & @xmath46 2.0 + 2.5 @xmath10 0.6 & + 12.2 & 1.1 & & @xmath46 2.1 & & & & @xmath46 1.2 & & & & @xmath46 1.0 & & & @xmath46 4.0 + + 2.9 @xmath10 0.4 & + 3.2 & 1.1 & & @xmath46 1.2 & & & & @xmath46 1.0 & & & & 3.0 @xmath10 0.2 & + 3.9 & 1.5 & 10.3 @xmath10 1.6 + 10.9 @xmath10 0.4 & + 6.8 & 1.6 & & 2.8 @xmath10 0.4 & + 6.5 & 1.5 & & @xmath46 1.0 & & & & 6.8 @xmath10 0.4 & + 7.1 & 0.8 & 6.2 @xmath10 0.4 + 1.8 @xmath10 0.6 & + 12.6 & 2.3 & & @xmath46 1.2 & & & & @xmath46 1.0 & & & & @xmath46 1.0 & & & @xmath46 5.6 + + 4.9 @xmath10 0.7 & @xmath53 & 1.8 & & @xmath46 1.3 & & & & @xmath46 1.5 & & & & 0.3 @xmath10 0.04 & @xmath293.7 & 0.5 & 0.6 @xmath100.1 + 2.6 @xmath10 0.6 & + 3.3 & 1.5 & & @xmath46 1.3 & & & & @xmath46 1.5 & & & & 3.1 @xmath10 0.8 & + 4.8 & 2.7 & 11.9 @xmath10 4.1 + 9.0 @xmath10 0.5 & + 6.2 & 1.3 & & @xmath46 1.3 & & & & @xmath46 1.5 & & & & 7.5 @xmath10 0.8 & + 6.7 & 1.1 & 8.3 @xmath10 1.0 + 2.0 @xmath10 0.4 & + 9.4 & 0.4 & & @xmath46 1.3 &
& & & @xmath46 1.5 & & & & @xmath46 1.0 & & & @xmath46 0.5 + 0.9 @xmath10 0.2 & + 12.0 & & & @xmath46 1.3 & & & & @xmath46 1.5 & & & & @xmath46 1.0 & & & @xmath46 1.1 + + 1.4 @xmath10 0.7 & + 5.0 & 1.8 & & @xmath46 2.1 & & & & @xmath46 1.3 & & & & @xmath46 1.7 & & & @xmath46 12.1 + 8.6 @xmath10 0.8 & + 7.2 & 1.6 & & 4.7 @xmath10 0.7 & + 7.5 & 2.2 & & @xmath46 1.3 & & & & 9.7 @xmath10 0.6 & + 6.9 & 1.1 & 11.3 @xmath10 1.3 + an extensive comparison of our values of total @xmath39 with those found in the literature appears in table 2 for the five species studied .
a thorough examination reveals that our values , obtained by summing all of the individual components along a line of sight , are consistent with most previous determinations , many of which derive from only one resolved component . of the 37 spectra in our investigation with counterparts in the literature , significant discrepancies in total @xmath39 ( greater than 3-@xmath38 )
are noted in only 6 .
stellar contamination most likely affected the measurements by white ( 1984__a _ _ ) of toward 20 tau and 22 tau and of ch toward hd 23512 .
white s ( 1973 ) measurement of toward @xmath20 tau is also suspected to suffer from inadequate normalization of a stellar feature .
our value for @xmath39 ( ) toward 27 tau is substantially larger than the consistent values given by three previous investigations ( younan & dufton 1984 ; white 1984__a _ _ ; marschall & hobbs 1972 ) .
this is due to our ability to detect weaker features in the wings of the more prominent absorption line .
similarly , the determination of @xmath39(ch@xmath0 ) toward hd 23753 by frisch ( 1972 ) was affected by her detection limit of @xmath2 6 m . in the case of non - detections , since our upper limits are significantly lower than those of all preceding studies , our results further constrain the detectable amount of absorption from these less abundant species along the observed sight lines .
we also point out that while federman ( 1982 ) claimed a detection of ch toward 20 tau , it is but a 2.5-@xmath38 detection and is consistent with our quoted upper limit .
the average uncertainties in our measurements of @xmath39 for individual components fall below 1 m for all four detected species . specifically , @xmath57 = 0.7 m for cn , 0.4 m for , and 0.3 m for ch@xmath0 and ch .
since the uncertainties from the literature are usually greater than 1 m , our measurements are among the most precise now available .
column densities for individual components were interpolated from curves of growth based on the measured equivalent widths and the parameters listed in table 3 .
the adopted @xmath1-values , 1.6 km s@xmath3 for , ch@xmath35 , and ch and 0.5 km s@xmath3 for cn , are average values for the respective species .
however , most lines are weak so that measured equivalent widths for a given species fall on the linear portion of the curve of growth .
thus , the use of different @xmath1-values does not impact the derived column densities in any appreciable way .
in particular , the largest optical depth at line center encountered under our assumptions is @xmath58 for the r(0 ) line of cn toward hd 23512 .
table 4 summarizes our results for the individual components of k , ch@xmath35 , and ch , listing the column density , @xmath59 , radial velocity , @xmath4 , and @xmath1-value of each component .
the reported uncertainty in @xmath59 for a given component was inferred from the measured uncertainty in the value of @xmath39 .
table 4 also displays the relevant parameters of the components found by white et al .
( 2001 ) in their survey of interstellar in the pleiades .
the last column gives the / column density ratio which will be discussed further in 4.2 .
for compactness , the cn column densities toward hd 23512 are not listed in table 4 but are given here .
@xmath59(cn ) = 2.7 @xmath10 0.3 , 1.6 @xmath10 0.3 , and @xmath46 2.8 in units of @xmath60 @xmath61 for the r(0 ) , r(1 ) , and p(1 ) lines , respectively .
these quantities will be important later for deriving the density of the molecular cloud along this line of sight .
lcccc ca ii & 20 & 65 & + 6.2 & 1.5 @xmath10 0.4 + & 20 & 55 & + 7.1 & 1.5 @xmath10 0.4 + ch@xmath35 & 19 & 23 & + 8.6 & 1.8 @xmath10 0.3 + ch & 6 & 6 & + 8.7 & 1.6 @xmath10 0.1 + cn & 1 & 1 & + 9.4 & 0.5 @xmath10 0.1 + na i & 20 & 42 & + 7.9 & 1.2 @xmath10 0.8 + & 20 & 37 & + 8.0 & 1.0 @xmath10 0.3 + the mean and dispersion in @xmath1-values were calculated for the detected species of this survey as well as for from the data provided by white et al .
the results are displayed in table 5 . to facilitate a meaningful comparison , the values in the last row of the table
were calculated by excluding five broad components with @xmath62 km s@xmath3 .
the scatter in @xmath1-values for the species in our investigation is low and is comparable to the scatter found for the narrow components .
our mean @xmath1-values show good agreement with those from previous investigations , both within the pleiades and in other regions of the ism , for which spectra of comparable snr were acquired .
the ultra - high resolution survey of ch@xmath0 and ch by crane et al .
( 1995 ) , for instance , included four pleiades stars , 20 tau , 23 tau , @xmath20 tau , and 27 tau .
while they did not claim to detect ch along these sight lines , their mean ch@xmath0 @xmath1-value for the four stars ( 1.9 @xmath10 0.5 km s@xmath3 ) agrees closely with the mean value for our sample of nineteen sight lines ( 1.8 @xmath10 0.3 km s@xmath3 ) .
welty et al . ( 1996 ) observed two pleiades stars , 17 tau and @xmath20 tau , in a high - resolution survey of .
their average @xmath1-value for the six components detected toward these two stars ( 1.5 @xmath10 0.9 km s@xmath3 ) is identical to our average for twenty stars ( 1.5 @xmath10 0.4 km s@xmath3 ) , though the median value for their entire sample is somewhat smaller ( @xmath2 1.3 km s@xmath3 ) .
pan et al .
( 2005 ) measured average @xmath1-values for a number of atomic and molecular species in an extensive examination of interstellar absorption in three star - forming regions , @xmath63 oph , cep ob2 , and cep ob3 . based on detections of 350 total components for and over a hundred components for the other species excluding cn for which they detect 50 components , these authors found average @xmath1-values of 1.62 @xmath10 0.28 km s@xmath3 for , 1.96 @xmath10 0.23 km s@xmath3 for ch@xmath0 , 1.04 @xmath10 0.18 km s@xmath3 for ch , and 0.83 @xmath10 0.11 km s@xmath3 for cn .
given the uncertainties , the mean @xmath1-values in table 5 are consistent with these results , except in the case of ch .
the larger value we obtain for this species ( 1.6 @xmath10 0.1 km s@xmath3 ) may have implications for the environment responsible for the formation of these lines .
previous studies have shown that ch can exist in both low - density gas ( @xmath64 @xmath7 ) where it is produced via non - thermal ch@xmath0 chemistry ( draine & katz 1986 ; zsarg and federman 2003 ) , and in higher - density regions ( @xmath65 @xmath7 ) where cn resides ( federman et al .
ch components associated with ch@xmath0 can be identified by a broad `` ch@xmath0-like '' absorption profile , whereas those found in regions where cn is detected will exhibit a sharper `` cn - like '' profile ( lambert , sheffer , & crane 1990 ) . in the analysis of pan et al .
( 2005 ) , cn - like ch components were discovered to have an average @xmath1-value indistinguishable from that of cn components ( 0.90 @xmath10 0.11 km s@xmath3 ) , while ch@xmath0-like ch components had a larger average @xmath1-value ( 1.10 @xmath10 0.16 km s@xmath3 ) .
the ch components examined in the pleiades are more closely associated with ch@xmath0 than along average sight lines .
indeed , our mean ch @xmath1-value is nearly indistinguishable from our mean value for ch@xmath0 and well above the @xmath1-values of the only detected cn lines ( 0.5 and 0.6 km s@xmath3 for r(0 ) and r(1 ) , respectively , toward hd 23512 ) . toward most of the pleiades , therefore , ch is found to be ch@xmath0-like as opposed to cn - like .
that is , ch is linked to the formation of ch@xmath0 rather than cn and traces lower - density interstellar clouds .
this conclusion would not be altered had we considered @xmath40-doubling in the ch line , which would lower our ch @xmath1-values by @xmath66 km s@xmath3 .
table 5 also presents the mean radial velocities of the various species weighted by the column density of each component .
these quantities yield information on the bulk kinematic properties of the various constituents of the interstellar gas near the pleiades .
the column - density weighted mean velocity for in our sample is @xmath2 + 6 km s@xmath3 , whereas for molecular species this average is closer to @xmath2 + 9 km s@xmath3 . if only positive velocity components are included , the mean velocity increases to @xmath2 + 7 km s@xmath3 .
white ( 1984__a _ _ ) was first to note the kinematic separation of atomic and molecular gas in this cluster .
his results yielded average velocities of @xmath2 + 7 km s@xmath3 for and @xmath2 + 8 km s@xmath3 for ch@xmath0 .
the distinction may even be more pronounced in the average velocities reported here especially considering our measurements of ch and cn velocities .
the components ( white et al . 2001 ) have a weighted mean velocity of @xmath2 + 8 km s@xmath3 in somewhat closer agreement with the molecular species than with , suggesting that ionized and neutral gas may also be kinematically distinct . in the following section ,
the bulk motions described above are examined in greater detail by identifying individual velocity components present in the atomic and molecular gas .
in order to analyze the complex interactions between the interstellar gas and the stars of the pleiades and to attempt to map the extent of the clouds involved in these interactions , contour maps were generated showing the spatial distribution of individual velocity components for species with detectable absorption along more than one line of sight .
these maps are presented in figures 6@xmath298 for ch , ch@xmath0 , and in order of increasing complexity .
the contours in these figures were drawn by eye to include all sight lines which exhibit a given component . however , since we only have information at a limited number of discrete points , unlike radio data for instance , the boundaries shown are approximate at best . as a rule ,
if a component was found along one sight line but not along another adjacent to it , a boundary was drawn half way between .
when no sight line was available to constrain the outer portions of a given component , contours were drawn to the edge of the plot . for the molecular species , ch and ch@xmath0 ,
one component was typically found per absorbing sight line , falling in either of two categories , those with @xmath4 @xmath2 + 7 km s@xmath3 and those with @xmath4
@xmath2 + 9.5 km s@xmath3 .
the former correspond to the `` central '' component in the profiles analyzed by white ( 2003 ) who identifies this absorption with the pervasive foreground gas thought to be associated with the taurus dust clouds ( see also white & bally 1993 ) .
the latter , also detected in the data , are termed the `` red '' component by white ( 2003 ) who suggests that this is either taurus gas redshifted to higher radial velocities by the interaction with cluster stars or an independent gas cloud also interacting with the pleiades . in our data , the strongest molecular absorption , including the only absorption detected in cn , is associated with the higher - velocity component .
the lower - velocity gas is more pervasive and the two components generally occupy different regions of the cluster .
figure 6 clearly illustrates the spatial distinction between 7 km s@xmath3 and 9.5 km s@xmath3 gas in the velocity components of ch and demonstrates the trend in absorption strength as well .
there are five weak components with a mean velocity of + 6.9 km s@xmath3 concentrated in the central region of the cluster .
four of these have column densities below @xmath67 @xmath61 and the fifth , toward hd 23568 , has a column density of @xmath68 @xmath61 .
a much stronger component , with @xmath69 @xmath61 , is detected at a velocity of + 9.8 km s@xmath3 toward hd 23512 .
the sight line to hd 23512 passes through a small molecular cloud first mapped in co emission by r.s .
cohen ( 1975 , unpublished ; see federman & willson 1984 ) .
the asterisk in figure 6 marks the position of peak co emission from the cloud .
federman and willson published cohen s co data and added additional radio observations of co and @xmath70co , as well as ch and oh , determining the velocity of the cloud to be about 10 km s@xmath3 . our velocity for the ch component toward hd 23512 is consistent with their determination .
in addition , the r(0 ) and r(1 ) lines of cn detected toward this star give velocities of + 9.4 and + 9.5 km s@xmath3 , respectively .
cn traces dense clouds and the extensive study by pan et al .
( 2005 ) found that strong ch absorption is associated with cn ( see 3.3 ) .
thus , it is not surprising that cn is only found along a sight line containing a molecular cloud and that the cn - like ch component in this direction has a consistent velocity .
although the ch@xmath0 components in figure 7 exhibit more complexity in their spatial distribution , the general trends are still apparent .
fourteen components are detected at a mean velocity of + 7.3 km s@xmath3 .
these are distributed in mainly the central and eastern parts of the cluster , with some components extending to the south and to the west .
the majority have column densities of @xmath71@xmath29@xmath72 @xmath61 .
the higher - velocity components responsible for the largest column densities of ch@xmath0 are concentrated in the western half of the cluster .
nine components are found at a mean velocity of + 9.3 km s@xmath3 and six of these have column densities greater than @xmath73 @xmath61 .
the components toward 17 tau and 23 tau have the highest velocities of this group , + 10.3 and + 10.2 km s@xmath3 , respectively , noticeably higher than the velocity of any other component .
these velocities coincide with the velocity of the molecular cloud discussed above and the sight lines toward these stars lie along the cloud s outer edge as seen in molecular emission ( federman & willson 1984 )
. both are strong indications that the material responsible for the absorption originated in the molecular cloud .
interestingly , no 7 km s@xmath3 gas is detected in the cloud s immediate vicinity ( toward hd 23512 ) but is seen in weak absorption to the north ( toward 17 tau and 23 tau ) and in more moderate absorption to the south ( toward hd 23410 ) .
we do not detect any systematic velocity shifts between our ch and ch@xmath0 lines .
the average difference in velocity for components detected in both species is @xmath42(ch@xmath43(ch@xmath74 km s@xmath3 .
our ch velocities also agree well with those of , with an average difference for individual sight lines of @xmath42(ch@xmath75(@xmath76 km s@xmath3 .
while the molecular data can be adequately represented by two distinct velocity components , the situation for is considerably more complex .
up to five separate components are identified along a single line of sight . in all , 55 components are detected with velocities distributed almost continuously from + 3.1 to + 14.0 km s@xmath3 .
another 10 are found with velocities between @xmath298.2 and @xmath293.7 km s@xmath3 .
velocities alone are not enough to categorize components with such diversity .
thus , the panels displayed in figure 8 were created by classifying the components not only according to their velocity , but also to their proximity to like sight lines and their overall position in the cluster .
the notable features of this figure are as follows .
the most pervasive group of components ( figure 8__e _ _ ) has a mean velocity of + 6.8 km s@xmath3 , clearly associating it with the central component of white s ( 2003 ) analysis and that found in our molecular data .
this component pervades the observed extent of the cluster with the exception of five sight lines along the northern edge , 16 tau , 19 tau , 21 tau , hd 23568 , and hd 23873 . unlike the weaker molecular feature at this velocity ,
the central component is the dominant source of atomic absorption in the pleiades with column densities ranging from @xmath77@xmath29@xmath78 @xmath61 .
proceeding outward in velocity from the central component , there are two groups ( figures 8__d _ _ and 8__f _ _ ) which appear nearly identical in their morphology and spatial extent .
these groups have mean velocities of + 5.5 and + 8.2 km s@xmath3 , respectively , and seem to populate the apparent void in the central component .
indeed , all five of the aforementioned northern sight lines have components at both the slightly higher and slightly lower velocities .
we interpret this bifurcation in velocity components , which appears smooth at our resolution , as evidence of a velocity gradient in the pervasive central cloud .
however , the number of components belonging to groups 8__d _ _ and 8__f _ _ may be complicated by unresolved component blending in the central component ( 8__e _ _ ) .
for example , welty et al . ( 1996 ) fit the asymmetric central region of their higher - resolution profile toward @xmath20 tau with two components with velocities of + 6.2 and + 7.7 km s@xmath3 , compared to our one component with a velocity of + 7.2 km s@xmath3 . in our data ,
a hint of this asymmetry in the central region of the profile can be seen for @xmath20 tau , as well as for 27 tau and 28 tau . in general , though , the @xmath1-values for these three groups of components are quite similar ( @xmath79 km s@xmath3 ) and thus do not indicate that unresolved structure is a major concern . moving out still further in velocity from the central component , we encounter two additional groups ( figures 8__c _ _ and 8__g _ _ ) with analogous spatial characteristics .
these groups , with mean velocities of + 3.8 and + 9.4 km s@xmath3 , respectively , exhibit identical structure in the cluster s northwest quadrant , but the lower - velocity components ( 8__c _ _ ) are more extensive in the southeast .
interestingly , the six sight lines that are missing from the feature in figure 8__g _ _ have components at even higher velocities ( figure 8__h _ _ ) . this group ( 8__h _ _ ) has a mean velocity of + 11.5 km s@xmath3 . taken together , the components in figures 8__g _ _ and 8__h _ _ are roughly comparable to the red component in white s ( 2003 ) classification scheme . those in figure 8__c _ _ correspond to what white ( 2003 ) terms the `` blue '' component , with the implication that gas from the central cloud at @xmath4 @xmath2 + 7 km s@xmath3 has been blueshifted by its interaction with the cluster to lower radial velocities .
the spatial correlations and morphological similarities of redshifted and corresponding blueshifted components in figure 8 are striking and support the claim that velocity gradients reflect the interaction on a cloud .
furthermore , the mean velocities found for the components of figures 8__c _ _ and 8__g _ _ differ from the mean velocities of the previous two groups ( figures 8__d _ _ and 8__f _ _ ) by an amount consistent with the velocity differences between those groups and the central component .
in other words , figures 8__c _ _ through 8__g _ _ present successive velocity intervals of approximately 1.4 km s@xmath3 evincing the symmetry of the interstellar gas in the pleiades traced by absorption .
the remaining panels of figure 8 present further components which do not conform to the above symmetries .
the components with the most extreme positive velocities ( figure 8__i _ _ ) inhabit a region directly west of the cluster center .
it is notable that the two sightlines showing this component , 17 tau and 23 tau with velocities of + 14.0 and + 13.7 km s@xmath3 , respectively , are the same as those with the highest ch@xmath0 velocities , although the actual velocities do not agree .
the proximity of these highly redshifted components to the edge of the molecular cloud is suggestive of strong interactions between cloud material and the luminous blue giants .
figures 8__a _ _ and 8__b _ _ reveal the distribution of absorption detected at negative velocities .
the six components concentrated in the central region of the cluster ( 8__b _ _ ) have a mean velocity of @xmath294.1 km s@xmath3 in close agreement with the negative - velocity components found in .
white ( 2003 ) names this the `` shocked '' component , referring to the prediction that strongly blueshifted absorption features would arise in foreground gas shocked by the interaction with cluster stars .
four additional shocked components are detected along the cluster s western edge ( 8__a _ _ ) .
these have even more strongly blueshifted velocities with a mean of @xmath296.3 km s@xmath3 .
white finds shocked components toward only hd 23410 , @xmath20 tau , 27 tau , and hd 23964 yet hints at unpublished data confirming additional detections toward 21 tau and 28 tau .
we also detect shocked components toward 21 tau and 28 tau and add to this list 16 tau , 19 tau , 23 tau , and 24 tau .
the shocked atomic components provide an important observational constraint when considering the types of interactions needed to explain the features of the interstellar medium in the pleiades .
, comparable to our velocity resolution.,scaledwidth=40.0% ] we performed a statistical analysis to determine whether binning the velocity components in figure 8 differently would severely alter our results . to this end , a series of histograms were constructed with successive bin sizes from 0.2 km s@xmath3 to 1.9 km s@xmath3 . in almost every case
, four distinct peaks could be seen for the positive - velocity components , with another single peak evident for the components with negative velocity .
a representative histogram is shown in figure 9 with a bin size of 1.4 km s@xmath3 , comparable to the resolution in our data .
the peaks in the figure are centered about values of @xmath294.5 , + 3.9 , + 6.7 , + 9.5 , and + 12.3 km s@xmath3 , closely echoing the mean velocities found for the components in figures 8__b _ _ , 8__c _ _ ,
8__e _ _ , 8__g _ _ , and 8__h _ _ , respectively . because variations in bin size have a negligible effect on the results presented in figure 8 , these features are not an artifact of a particular classification scheme , but represent physical structure in the interstellar atomic gas .
the last column of table 4 displays the @xmath59()/@xmath59 ( ) ratios for our components common to the investigation of white et al .
( 2001 ) and upper limits for components not detected in .
figure 10 plots these ratios as a function of @xmath42 ( ) .
routly and spitzer ( 1952 ) were first to demonstrate that the / column density ratios are inversely proportional to the velocity of the material .
this phenomenon , known as the routly - spitzer effect , was later confirmed by siluk and silk ( 1974 ) , who found that the decrease in @xmath59()/@xmath59 ( ) was most evident for components with absolute velocities greater than @xmath2 20 km s@xmath3 .
the routly - spitzer effect is generally attributed to variations in the gas - phase abundance of calcium in clouds at different velocities .
calcium is depleted onto the surfaces of dust grains in low - velocity diffuse interstellar clouds resulting in high / ratios in these environments .
heating from shock waves and cloud collisions in higher - velocity material can cause grain destruction which releases calcium back to its gaseous phase , yielding a lower / ratio .
the data in figure 10 show an indication of this trend but there are not enough high - velocity detections to be certain .
( na i)/@xmath59(ca ii ) ratios as a function of @xmath42(ca ii ) . of the four data points with @xmath80 ,
three have ratios less than one .
these are for the shocked components toward @xmath20 tau , 27 tau , and hd 23964 .
the fourth , with a ratio of @xmath2 2 , is for the shocked component toward hd 23410 .
the lower limit near the top of the figure is for the + 9.2 km s@xmath3 component toward hd 23512.,scaledwidth=40.0% ] more importantly , because calcium depletion is assumed to be density - dependent , our / ratios can be used to characterize the physical conditions of the interstellar clouds responsible for the detected absorption , providing a test for our derived column densities .
the majority of positive - velocity components in figure 10 have @xmath59()/@xmath59 ( ) of order a few to 10 , consistent with the values typical of diffuse , non - molecular clouds with densities of @xmath81 @xmath7 ( hobbs 1976 ) .
the only exception to this is the + 9.2 km s@xmath3 component toward hd 23512 .
the / ratio for this component ( the lower limit near the top of figure 10 ) is of order 100 , more characteristic of a denser , molecular cloud where calcium is depleted onto grain surfaces more effectively ( crawford et al .
1989 ; crawford 1992 ) .
in contrast , the + 6.2 km s@xmath3 component along this sight line has a ratio of @xmath2 10 placing it among the diffuse components where calcium depletion is less severe . since these results predict the proper relationship between velocity components and the expected physical characteristics of the associated clouds , it gives us confidence in the accuracy of our derived column densities .
a conspicuous feature of figure 10 is that the / ratios for shocked atomic components ( those with @xmath82 ) are less than one except toward hd 23410 for which @xmath59()/@xmath59 ( ) @xmath2 2 . in the environment near the pleiades , a ratio less than one probably implies that photoionization from bright cluster stars has led to a reduction in neutral sodium .
this effect would be less pronounced for gas near hd 23410 lying far from the more luminous stars and producing little luminosity of its own .
the shocks themselves do not have high enough velocities to contribute much gas - phase calcium from grain processing .
it is possible to extract the physical conditions of the interstellar medium near the pleiades directly from our absorption - line data . in this section ,
we derive estimates for @xmath83 , the total gas density of the absorbing clouds , and @xmath84 , the enhancement of the uv radiation field due to cluster stars over the average galactic field , by adopting various models of diffuse - cloud chemistry . for the majority of sight lines in our sample , ch is found to be closely associated with ch@xmath0 . in such environments , the total gas density , defined as @xmath85(h ) + @xmath86(h@xmath87 ) , may be obtained by studying the reactions which lead to the production of ch from ch@xmath0 . in the calculations that follow ,
we do not explicitly consider the factors responsible for ch@xmath0 formation .
instead , we follow the chemistry once ch@xmath0 is present in appreciable amounts . in particular
, we use the observed column density of ch@xmath0 to determine the gas density implied by the amount of ch present . assuming that the most important reactions @xmath88 satisfy the conditions for a steady state , the rate equation for ch may be written ( welty et al .
2006 ) @xmath89 where the factor 0.67 represents the fraction of dissociative recombinations of ch@xmath90 producing ch ( herbst 1978 ; vejby - christensen et al .
1997 ) , @xmath91(ch@xmath0,h@xmath92 @xmath93 s@xmath3 is the rate coefficient for the initial reaction ( draine & katz 1986 ) , @xmath94(h@xmath95(h@xmath5)/@xmath96(h ) is the fraction of hydrogen in molecular form , and @xmath97(ch@xmath98 s@xmath3 is the ch photodissociation rate for the average galactic field ( federman & huntress 1989 ) . in the expression for @xmath94(h@xmath5 ) , @xmath96(h ) = @xmath59 ( ) + @xmath99(h@xmath5 ) denotes the total hydrogen column density along the line of sight .
( 1 ) may be inverted to derive an expression for the gas density as a function of the ch / ch@xmath0 column density ratio , @xmath100 lccccccccccc 16 tau & 0.62 & @xmath101 0.3 & 28.9 & @xmath101 168 & & & & & & + & & @xmath46 0.06 & & @xmath46 28 & & & & & & + 17 tau & 0.31 & @xmath46 0.8 & 21.9 & @xmath46 404 & & & & & & + & & @xmath46 0.5 & & @xmath46 280 & & & & & & + 18 tau & 0.31 & @xmath46 0.3 & 13.3 & @xmath46 88 & & & & & & + 19 tau & 0.25 & @xmath46 0.2 & 34.5 & @xmath46 158 & & & & & & + 20 tau & 0.43 & @xmath46 0.05 & 14.9 & @xmath46 16 & & 2.4 @xmath10 1.0 & 3.2 @xmath10 1.3 & 1.3 @xmath10 0.2 & 89 @xmath10 34 & 8.6 @xmath10 3.4 + & & @xmath46 0.05 & & @xmath46 16 & & & & & & + 21 tau & 0.43 & @xmath46 0.3 & 14.7 & @xmath46 82 & & & & & & + & & @xmath46 0.07 & & @xmath46 21 & & & & & & + 22 tau & 0.37 & @xmath46 0.06 & 7.9 & @xmath46 10 & & & & & & + 23 tau & 0.62 & @xmath46 0.5 & 20.5 & @xmath46 192 & & 7.8 @xmath10 2.0 & 5.2 @xmath10 1.3 & 2.2 @xmath10 0.2 & 48 @xmath10 12 & 7.1 @xmath10 1.8 + & & @xmath46 0.04 & & @xmath46 14 & & & & & & + hd 23568 & 0.43 & @xmath101 0.5 & 6.4 & @xmath101 72 & & & & & & + & & @xmath46 0.06 & & @xmath46 8 & & & & & &
+ @xmath20 tau & 0.25 & 0.2 & 8.8 & 46 @xmath10 9 & & 1.6 @xmath10 0.4 & 1.9 @xmath10 0.5 & 0.8 @xmath10 0.1 & 83 @xmath10 20 & 6.3 @xmath10 1.6 + hd 23753 & 0.25 & @xmath46 0.2 & 11.0 & @xmath46 53 & & & & & & + 27 tau & 0.25 & 0.1 & 6.1 & 16 @xmath10 6 & & & & & & + 28 tau & 0.19 & 0.1 & 10.3 & 40 @xmath10 6 & & & & & & + hd 23923 & 0.37 & @xmath46 0.4 & 3.6 & @xmath46 29 & & & & & & + to estimate the gas density from eqn . ( 2 ) , a few important parameters must first be determined , namely @xmath94(h@xmath5 ) and @xmath84 .
the first of these , the molecular fraction , is not well constrained in the vicinity of the pleiades due to a dearth of reliable measurements of @xmath59 ( ) . as a starting point in our analysis , we adopted a value of @xmath94(h@xmath102 due to the strong presence of ch@xmath0 in these clouds .
a greater fraction of hydrogen in molecular form would rapidly destroy ch@xmath0 by the reaction ch@xmath0 + h@xmath5 @xmath8 ch@xmath103 + h. since @xmath83 is inversely proportional to @xmath94 , a lower value of @xmath94 would increase the densities obtained from eqn .
( 2 ) , a fact which is exploited in 4.3.2 .
individual values of @xmath84 are calculated using the following expression modified from eqn .
( 2 ) of white ( 1984__b _ _ ) , @xmath104\mathrm{e}^{-\tau_{uv}}+ \chi_{\lambda , c},\ ] ] where @xmath105 is the ratio of the flux due to a specific star at a distance of 1 pc to that produced by the unattenuated galactic background , @xmath106 is the corresponding ratio for the flux due to the other cluster stars at that position , and @xmath107 is the distance from the star in pc . the radiative parameters for 14 of our 20 stars ( at @xmath108 ) are obtained from table 1 of white ( 1984__b _ _ ) from which @xmath107 is taken to be @xmath109 , which is the minimum distance at which radiation pressure is balanced by the drag force on dust grains slipping through the gas as it approaches the cluster . yet , since @xmath107 in eqn . ( 3 ) refers to the gas responsible for the absorption , the adoption of these values assumes a close coupling between the gas and dust , though the actual distances used ( @xmath110@xmath29@xmath111 pc ) are quite similar to those found in other analyses . for example , jura ( 1977 ) argued for a distance of 0.1 pc for the gas toward 20 tau and federman ( 1982 ) determined the distance to be 0.3 pc .
the value obtained from white ( 1984__b _ _ ) for the gas toward this star is also 0.3 pc . finally , to calculate @xmath84 as well as @xmath97(ch ) , we approximate @xmath112 , the grain optical depth at 1000 , by @xmath113 ( federman et al .
1994 ) , determined from the reddening values in table 1 .
table 6 displays the relevant parameters for each sight line and the resulting density estimates . because the majority of sight lines show no detectable ch absorption , most of the densities in column ( 5 ) are only upper limits .
we do not attempt to apply this method to the sight line toward hd 23512 because the assumptions leading to eqn .
( 2 ) would not be appropriate for the gas in a molecular cloud where ch is linked to cn chemistry . of the remaining five sight lines with both ch@xmath0 and ch in detectable amounts , two ( 16
tau and hd 23568 ) do not have components at similar velocities .
thus , only three actual estimates for density can be obtained from our data using this procedure .
the values , @xmath83 = 46 , 16 , and 40 @xmath7 for the sight lines toward @xmath20 tau , 27 tau , and 28 tau , respectively , are significantly below the densities predicted by earlier investigations .
the atomic ionization models of white ( 1984__a _ _ ) favored a density of 400 @xmath7 , while his chemical models of ch and ch@xmath0 commonly employed values of 100 and 300 @xmath7 ( white 1984__b _ _ ) .
the 21 cm data examined by gordon & arny ( 1984 ) also suggested a hydrogen density of @xmath114 @xmath7 .
however , a recent analysis by zsarg and federman ( 2003 ) of excitation yielded a gas density upper limit toward @xmath20 tau , @xmath115 @xmath7 , an order of magnitude below our determination .
an alternative approach to obtaining the physical conditions of the diffuse clouds near the pleiades involves the distribution of h@xmath5 column densities in the @xmath116 = 0 , 1 , and 4 rotational levels of the ground vibrational state .
much of this analysis , adapted from lee et al .
( 2002 ) , is derived from the work of jura ( 1974 , 1975 ) who showed that the @xmath116 = 4 and 5 rotational levels of h@xmath5 are populated primarily by photon pumping . in environments of low to moderate density
, these levels will be depopulated by spontaneous emission .
thus , the relative populations of higher and lower rotational levels should indicate the density of the gas . while there may be other sources of excitation active in these clouds , such as collisions in shocked gas
, the present model considers an extreme case in which all rotational excitation is due to uv pumping . here
, again , we assume steady - state conditions and obtain an expression for the gas density from eqn .
( 3 ) of lee et al . ( 2002 ) adopting an h@xmath5 formation rate coefficient , @xmath117 @xmath93 s@xmath3 , appropriate for galactic clouds .
the expression is @xmath118}+0.19\right\}^{-1},\ ] ] where @xmath119 is the h@xmath5 column density in rotational level @xmath116 and the total hydrogen column density is found from @xmath96(h@xmath120(h@xmath121(h@xmath122/f$](h@xmath123 . in this case , therefore , @xmath83 is directly proportional to @xmath94 . column ( 9 ) of table 6 presents the density estimates derived from eqn .
( 4 ) for the three sight lines with h@xmath5 column densities measured by _ copernicus _
( spitzer , cochran , & hirshfeld 1974 ; savage et al . 1977 ; frisch & jura 1980 ) .
unfortunately , only one of the sight lines , that toward @xmath20 tau , has a corresponding value from the @xmath59(ch)/@xmath59(ch@xmath0 ) method
. nevertheless , the densities obtained from eqn . ( 4 ) , @xmath83 = 89 , 48 , and 83 @xmath7 for 20 tau , 23 tau , and @xmath20 tau , respectively , are consistently larger than those from the previous method by about a factor of two . here
we exploit the opposite dependence of @xmath83 on @xmath94 exhibited in eqns .
( 2 ) and ( 4 ) to constrain both the density and the molecular fraction .
table 7 shows the effect on @xmath83 calculated from the two methods as @xmath94 is decreased slightly from 0.1 to 0.05 .
the analysis finds that a value of @xmath124 provides the best agreement among the various density estimates , a value consistent with the average value of @xmath94 for sight lines in scorpius ( savage et al .
the density toward @xmath20 tau for this value of @xmath94 is then @xmath125 @xmath7 .
lccccccc 20 tau & @xmath46 16 & @xmath46 23 & @xmath46 32 & & 89 @xmath10 34 & 62 @xmath10 24 & 44 @xmath10 17 + & @xmath46 16 & @xmath46 22 & @xmath46 31 & & & & + 23 tau & @xmath46 192 & @xmath46 274 & @xmath46 384 & & 48 @xmath10 12 & 34 @xmath10 8 & 24 @xmath10 6 + & @xmath46 14 & @xmath46 19 & @xmath46 27 & & & &
+ @xmath20 tau & 46 @xmath10 9 & 66 @xmath10 14 & 92 @xmath10 19 & & 83 @xmath10 20 & 58 @xmath10 14 & 41 @xmath10 10 + 27 tau & 16 @xmath10 6 & 22 @xmath10 9 & 31 @xmath10 13 & & & & + 28 tau & 40 @xmath10 6 & 56 @xmath10 8 & 79 @xmath10 12 & & & &
+ lee et al .
( 2002 ) derive another expression which relates the gas density to the ambient uv radiation field .
we can use this relation to determine @xmath84 based on the h@xmath5 column densities in table 6 . from eqn .
( 4 ) of lee et al . ( 2002 ) and using the above value for @xmath126 , @xmath127^{1/2},\ ] ] where @xmath128 s@xmath3 is the h@xmath5 photoabsorption rate corresponding to the average galactic field and the factor 0.11 represents the fraction of absorptions leading to h@xmath5 dissociation .
the @xmath84 values calculated from eqn . ( 5 ) are given in the last column of table 6 .
while there are only three determinations , each is smaller by approximately a factor of two compared to the values in column ( 4 ) calculated from the parameters in white ( 1984__b _ _ ) .
however , since eqn .
( 5 ) has no explicit @xmath107-dependence , we can determine what value of @xmath107 brings our @xmath84 values into agreement . solving eqn .
( 3 ) for @xmath107 and using @xmath84 from column ( 10 ) of table 6 yields
@xmath107 = 0.9 , 1.1 , and 1.3 pc for 20 tau , 23 tau , and @xmath20 tau , respectively .
these are to be compared to the distances obtained from white ( 1984__b _ _ ) for the same three stars , @xmath107 = 0.30 , 0.32 , and 0.72 pc , the values initially used in eqn .
( 3 ) to calculate the @xmath84 values in column ( 4 ) of table 6 . by substituting eqn . ( 4 ) for @xmath83 into eqn . ( 5 ) , it can be shown that this formula for @xmath84 is also independent of @xmath94(h@xmath5 ) .
thus , if the larger distances just calculated were used in the comparative density analysis above , our best - match value of @xmath94 would decrease slightly to 0.06 , thereby decreasing the density toward @xmath20 tau to 52 @xmath7 . despite the generally good agreement found from the above chemical models , the solutions are not unique and the number of free parameters
certainly does not eliminate the need for further refinement .
one example would be the application of a time - dependent model of ch@xmath0 formation in this environment . assuming a flow rate for the gas with respect to the cluster of 10 km s@xmath3 ( gordon & arny 1984 ) and a cloud thickness of 0.1 pc ( white 1984__a _ _ ) , the timescale for cloud - cluster interactions is @xmath129 yrs .
a corresponding timescale for h@xmath5 dissociation can be approximated by 1/@xmath97(h@xmath5 ) , where @xmath97(h@xmath130 is the h@xmath5 photodissociation rate given above .
if self - shielding is also included , the dissociation rate would be lower by at least a factor of 10 yielding a characteristic timescale of @xmath131 yrs .
because the respective timescales are comparable , we can not decouple the two processes and time - dependent effects must ultimately be incorporated into any chemical model adopted to derive physical conditions . following the analysis of ch and cn chemistry by pan et al .
( 2005 ) , we obtained a value for the gas density of the pleiades molecular cloud toward hd 23512 .
the model is appropriate for a diffuse cloud with large molecular content but does not include dark - cloud chemistry . as argued by federman & willson ( 1984 ) , the co column density from the densest , most opaque portion of the cloud , @xmath59(co@xmath132 @xmath61 , while larger than the values measured in diffuse gas , is significantly lower than those measured toward the centers of dark clouds .
the pleiades cloud is not , therefore , an intermediate zone between diffuse and dark clouds but simply a much less opaque molecular cloud than is normally observed in the taurus region . using our measured ch and cn column densities toward hd 23512 and setting @xmath133 due to the position of this star almost a parsec distant ( in projected separation ) from the luminous cluster stars ,
the model suggests @xmath134 @xmath7 .
the original study by federman & willson ( 1984 ) predicted @xmath135 300@xmath29500 @xmath7 based on their @xmath70co data and the large velocity gradient model of goldsmith et al .
bally and white ( 1986 , unpublished ; see white 2003 ) found a density of 300 @xmath7 within 10@xmath136 of the position of peak co emission and 800 @xmath7 at the position of the peak .
the consistency of these densities with those derived by federman & willson ( 1984 ) leads us to conclude that our lower limit is in error , being a factor of two larger than even the above peak value . evidently , there is more cn along this line of sight than can be attributed solely to diffuse - cloud chemistry .
the chemical reaction network common to dark clouds appears to contribute to cn production despite the unusually low density of this diffuse molecular cloud .
in 4.1 , the relatively straightforward distribution of molecular gas in the pleiades as traced by the molecules of ch@xmath0 , ch , and cn is contrasted with the rather complicated structures seen in the ionized atomic gas traced by .
the largest columns of molecular gas are associated with a velocity component at @xmath4 @xmath2 + 9.5 km s@xmath3 concentrated in the western region of the cluster .
clearly , some of this material is related to the molecular cloud seen along the line of sight toward hd 23512 , particularly that detected in ch and cn absorption in this direction .
indeed , the velocities of the ch and cn components here are consistent with the cloud velocity determined from molecular line emission by federman & willson ( 1984 ) and , although the density of the cloud derived from our ch and cn column densities ( @xmath134 @xmath7 ) is significantly larger than their predicted value ( @xmath2 300@xmath29500 @xmath7 ) , the inclusion of dark - cloud chemistry should correct this discrepancy . still , whether or not all molecular components that share the cloud velocity have a common origin remains elusive .
the general conclusions of the expansive study by white ( 2003 ) favor a three body encounter in which gas associated with the taurus dust clouds and redshifted by the interaction with the pleiades accounts for some of the red components and a separate cloud approaching from the west and also interacting with the cluster accounts for others . in support of the latter ,
the strongest components of ch@xmath0 absorption are located northwest of the cluster center , with a position and orientation similar to that of the western cloud described by white ( 2003 ) .
the absence of any corresponding ch components in this region indicates the low density of the material .
gas toward 17 tau and 23 tau with a velocity of + 10 km s@xmath3 likely probes the low - density envelope of the molecular cloud which lies just to the south of these sight lines .
the rest of the ch@xmath0 components near + 9 km s@xmath3 may also trace this envelope , but the data do not preclude a connection with redshifted taurus material . the other , weaker constituent of molecular gas in the pleiades has a velocity of @xmath2 + 7 km s@xmath3 , associating it with the central component of atomic gas and with the taurus clouds .
all sight lines with detectable amounts of ch , except that toward hd 23512 , and the majority of ch@xmath0 sight lines exhibit this component .
it is found in both species along sight lines to three stars in the center of the cluster ( @xmath20 tau , 27 tau , and 28 tau ) and these show similar ch / ch@xmath0 column density ratios ( 0.2 for @xmath20 tau and 0.1 for 27 tau and 28 tau ) . toward hd 23512 , where both species are detected at @xmath2 + 9.5 km s@xmath3 ,
the ratio is much larger ( 0.9 ) . that the column densities of ch and ch@xmath0 are well correlated for the central component strengthens the claim that ch toward most of the pleiades is ch@xmath0-like .
the larger ratio toward hd 23512 results from the addition of cn - like ch to the total ch column and reinforces the idea that this sight line effectively probes the molecular cloud .
the patterns exhibited in the velocity components of ( figure 8) offer unique insight into the complex interactions between the ism and the stars of the pleiades .
for instance , the central component at @xmath4 @xmath2 + 7 km s@xmath3 seen toward all of the stars in our survey does not have a uniform velocity across the cluster .
the slight blueshifts and redshifts apparent in figures 8__d _ _ and 8__f _ _ , respectively , are a strong indication that a dynamical interaction is occurring in which cloud material passing through the uv radiation field of the cluster is deflected toward and away from our line of sight .
that the pattern seems to repeat itself at larger and smaller velocities ( figures 8__c _ _ and 8__g _ _ ) suggests that the encounter of this central component can account for at least some of the observed red and blue components .
gas detected at velocities higher than @xmath2 + 10 km s@xmath3 likely derives from multiple sources .
the feature seen in figure 8__h _ _ may contain components missing from the gas in figure 8__g _ _ which have been further redshifted by cluster interactions , but may also contain independent components from a chance meeting with a second approaching cloud .
the latter is the probable source of the extreme - velocity components in figure 8__i _ _ due to the proximity of these sight lines to the molecular cloud .
the shocked atomic components in figures 8__a _ _ and 8__b _ _ show additional indications of a velocity gradient from the center of the cluster to the northwestern edge , despite the noticeable gap toward 17 tau , 20 tau , and 22 tau .
in general , the various components of are more pervasive in number and extent than those in the analysis of observations by white ( 2003 ) , as expected if is the most widely distributed species as in figure 6 of pan et al .
clearly , highly sensitive observations of can trace the intricate cloud - cluster and cloud - cloud interactions of the pleiades in greater detail and on a larger scale than observations of equal quality .
an important result of our analysis of / ratios ( 4.2 ) was that they predicted the proper relationship between velocity components and the expected physical conditions of the associated clouds , strengthening our confidence in the column densities derived from our observations .
/ ratios of order a few to 10 were measured for most positive - velocity components indicating the diffuse nature of much of the interstellar gas near the cluster where calcium is not efficiently depleted onto the surfaces of dust grains . a ratio of order 100
was observed in the + 9.2 km s@xmath3 component toward hd 23512 since this gas lies in a denser region near the molecular cloud , an environment which results in higher calcium depletion .
final confirmation comes from the + 6.2 km s@xmath3 component toward this star whose / ratio of @xmath2 10 rightly places it among the diffuse components .
the precise density estimates of 4.3 can place meaningful constraints on the chemical processes active in the diffuse clouds of the pleiades .
a comparison of the values obtained from ch / ch@xmath0 column density ratios with those from rotationally excited h@xmath5 column densities offers particular insight .
we find generally good agreement in the densities obtained from the two methods by adopting a molecular fraction of @xmath124 .
this value of @xmath94 yields typical densities of @xmath137 @xmath7 , significantly below earlier predictions which favored densities between 100 @xmath7 ( gordon & arny 1984 ) and 400 @xmath7 ( white 1984__a _ _ , _ b _ ) .
a possible discrepancy with our more precise estimates comes from the recent study by zsarg and federman ( 2003 ) who inferred a density upper limit of @xmath115 @xmath7 from the column densities of the fine - structure levels of toward @xmath20 tau .
this determination is an order of magnitude lower than our value toward @xmath20 tau , @xmath125 @xmath7 , obtained from both ch / ch@xmath0 ratios and h@xmath5 rotational populations . yet , based on their density estimate , zsarg and federman ( 2003 ) predict a ch column density upper limit toward @xmath20 tau of @xmath59(ch ) @xmath138 @xmath61 , consistent with our measured value of @xmath139 @xmath61 . to derive this predicted column density , the authors incorporated non - thermal motions of ions and neutrals due to the propagation of alfvn waves in a transition zone between cloud and intercloud material ( federman et al .
1996__a _ _ ) . without incorporating turbulence
, the predicted column density along this sight line is @xmath59(ch ) @xmath140 @xmath61 , much lower than the measured value .
non - thermal models hold great promise for diffuse - cloud chemistry , particularly in explaining the abundance of the ch@xmath0 radical as well as the abundances of molecules like ch which are tied to its formation . the question remains , then , why the density predicted by zsarg and federman ( 2003 ) is so much lower than our value .
two possibilities immediately present themselves .
first , observations of h@xmath5 , ch , and ch@xmath0 may sample different cloud depths due to the stratification of molecular species along the line of sight .
this scheme suggests that rotationally excited h@xmath5 molecules probe denser portions of the cloud closer to the stars , in the main portion of the photodissociation region ( pdr ) , where the uv radiation is strongest and is responsible for populating the higher @xmath116 levels through uv pumping .
ch and ch@xmath0 , as well as , would then trace the less dense outer portions of the cloud away from the pdr .
a major drawback to this scenario is it requires @xmath141 or higher to yield low enough densities from our measured ch / ch@xmath0 column density ratios to agree with the result , which would subsequently increase our h@xmath5-derived densities for this diffuse gas to greater than 400 @xmath7 , a value more typical of the pleiades molecular cloud .
a more suitable alternative is that the fine - structure levels of trace regions of various extent along the line of sight .
if the @xmath142 level , for instance , was confined to a region of space 1/10 the size of that occupied by the @xmath143 level , the density upper limit would be more in line with our determination .
finally , the inclusion of time - dependent effects into a chemical model of ch@xmath0 formation may still prove to be important since the timescale for h@xmath5 dissociation is comparable to the characteristic time in which the cloud is interacting with the cluster . because the initiating reaction leading to ch production requires h@xmath5 , a decreasing presence due to photodissociation would demand a higher density to account for the observed columns of ch .
such a requirement could be accommodated by our simple steady - state model by , again , raising slightly the value of @xmath94(h@xmath5 ) .
the most general conclusion of this investigation is that detailed observations of both atomic and molecular species essentially confirm the scenario of cloud - cluster interactions constructed by white ( 2003 ) .
a pervasive foreground gas cloud at @xmath4 @xmath2 + 7 km s@xmath3 seen toward nearly every star in both strong atomic and weaker molecular absorption is flowing through the cluster and dynamically interacting with the stellar uv radiation field .
the clearly symmetric patterns of both blueshifted and redshifted components extending across the cluster demonstrate the large - scale nature of these interactions .
a second cloud at @xmath4 @xmath2 + 10 km s@xmath3 is also likely to be interacting with the cluster due to the strong ch@xmath0 absorption near this velocity northwest of the cluster center and the existence of a diffuse molecular cloud also at @xmath4 @xmath2 + 10 km s@xmath3 to the southwest .
high velocity components with no direct connection to redshifted foreground gas may also require the existence of a second interacting cloud .
twenty stars in the pleiades were observed with high resolution and high signal to noise in a spectral range allowing for the detection of absorption features from cn , k , , ch@xmath0 , and ch .
total equivalent widths are consistent with previous determinations except in cases where we detect weaker features owing to our greater sensitivity .
mean @xmath1-values for cn , ch@xmath0 , and ch indicate that toward most of the pleiades ch is linked to the formation of ch@xmath0 rather than to cn chemistry .
weighted mean velocities for individual species reveal a kinematic separation of atomic and molecular gas , with neutrals more closely associated with the latter . for the molecular species , two distinct velocity components are sufficient to characterize the observed profiles , with the sight line toward hd 23512 clearly probing the molecular cloud seen in this direction . considerably more structure is seen along lines of sight in the atomic gas traced by .
velocity gradients are apparent in the pervasive central component implying a dynamical interaction with the cluster .
density estimates from ch / ch@xmath0 column density ratios agree well with those based on the rotational excitation of h@xmath5 molecules for @xmath144h@xmath145 , yielding typical densities of @xmath137 @xmath7 for the ism near the pleiades .
however , given the similar timescales of the relevant chemical and dynamical processes , the importance of time - dependent effects must be examined . with the availability of high - quality absorption - line data for cn , , , ch@xmath0 , and ch provided by this investigation and for provided by white et al .
( 2001 ) , existing models of interstellar chemistry must now be improved and brought to bear on the complex interaction between diffuse gas and the pleiades cluster .
the precise densities determined here must be incorporated into any model describing the interaction and all phenomenon must be treated self - consistently .
the application of a time - dependent model , in particular , is likely to prove necessary .
we thank yaron sheffer for his help in synthesizing ch profiles and dan welty for his very useful comments on an earlier draft of this paper .
we also acknowledge the thorough examination of our calculations by the anonymous referee that led to more consistent results .
this research made use of the simbad database operated at cds , strasbourg , france .
m. martinez participated in the research experience for undergraduates at the university of toledo under nsf - reu grants phy-0097367 and 0353899 .
the research presented here was also supported in part by nasa long term space astrophysics grant nag5 - 4957 to the university of toledo .
crane , p. , lambert , d.l . , & sheffer , y. 1995 , , 99 , 107 crawford , i.a .
1992 , , 259 , 47 crawford , i.a . ,
barlow , m.j .
, & blades , j.c .
1989 , , 336 , 212 draine , b.t . , &
katz , n. 1986 , , 306 , 655 elitzur , m. , & watson , w.d .
1978 , , 222 , l141 elitzur , m. , & watson , w.d . 1980 , , 236 , 172 falgarone , e. , verstraete , l. , pineau des forts , g. , & hily - blant , p. 2005
, , 433 , 997 federman , s.r .
1982 , , 257 , 125 federman , s.r . , &
huntress , w.t .
1989 , , 338 , 140 federman , s.r . , rawlings , j.m.c . ,
taylor , s.d . , & williams , d.a .
1996__b _ _ , , 279 , l41 federman , s.r . , strom , c.j . , lambert , d.l . ,
cardelli , j.a . ,
smith , v.v . , & joseph , c.l .
1994 , , 424 , 772 federman , s.r . ,
weber , j. , & lambert , d.l .
1996__a _ _ , , 463 , 181 federman , s.r . , & willson , r.f .
1984 , , 283 , 626 frisch , p.c .
1972 , , 172 , 301 frisch , p.c . , &
jura , m. 1980 , , 242 , 560 gredel , r. , van dishoeck , e.f .
, & black , j.h 1993 , , 269 , 477 goldsmith , p.f . ,
young , j.s .
, & langer , w.d .
1983 , , 51 , 203 gordon , k.j . , &
arny , t.t .
1984 , , 89 , 672 hawkins , i. , & jura , m. 1987 , , 317 , 926 herbst , e. 1978 , , 222 , 508 hobbs , l.m .
1973 , , 181 , 79 hobbs , l.m .
1976 , , 206 , l117 joulain , k. , falgarone , e. , pineau des forts , g. , & flower , d. 1998 , , 340 , 241 jura , m. 1974 , , 191 , 375 jura , m. 1975 , , 197 , 581 jura , m. 1977 , , 218 , 749 lambert , d.l . , sheffer , y. , & crane , p. 1990
, , 359 , l19 lee , d .- h . ,
min , k .- w . , federman , s.r .
, ryu , k .- s .
, han , w .- y . , nam , u .- w . , chung , h .- s . ,
dixon , w.v.d . ,
& hurwitz , m. 2002 , , 575 , 234 marschall , l.a . , &
hobbs , l.m .
1972 , , 173 , 43 morton , d.c . 1991 , ,
77 , 119 pan , k. , federman , s.r . ,
cunha , k. , smith , v.v .
, & welty , d.e .
2004 , , 151 , 313 pan , k. , federman , s.r . , sheffer , y. , & andersson , b .-
2005 , , 633 , 986 perryman , m.a.c . , ed .
1997 , the _ hipparcos _ and tycho catalogues ( esa sp-1200 ; noordwijk : esa ) pineau des forts , g. , flower , d.r . , hartquist , t.w . , & dalgarno , a. 1986 , , 220 , 801 routly , p.m. , & spitzer , l. , jr .
1952 , , 115 , 227 savage , b.d . ,
bohlin , r.c .
, drake , j.f . , & budich , w. 1977 , , 216 , 291 siluk , r.s .
, & silk , j. 1974 , , 192 , 51 spitzer , l. , jr . , cochran , w.d . , & hirshfeld , a. 1974 , , 28 , 373 tull , r.g . , macqueen , p.j . , sneden , c. , & lambert , d.l .
1995 , , 107 , 251 vallerga , j.v . , vedder , p.w . , craig , n. , & welsh , b.y .
1993 , , 411 , 729 vanden bout , p.a . , & snell , r.l . 1980 , , 236 , 460 vanden bout , p.a . , &
thaddeus , p. 1971 , , 170 , 297 vejby - christensen , l. , andersen , l.h . , heber , o. , kella , d. , pedersen , h.b .
, schmidt , h.t . , & zajfman , d. 1997 , , 483 , 531 welty , d.e . ,
federman , s.r . ,
gredel , r. , lambert , d.l . , & thorburn , j.a .
2006 , submitted to welty , d.e . ,
hobbs , l.m .
, & morton , d.c .
2003 , , 147 , 61 welty , d.e . ,
morton , d.c . , & hobbs , l.m .
1996 , , 106 , 533 white , r.e . 1973 , ,
183 , 81 white , r.e .
1984__a _ _ , , 284 , 685 white , r.e .
1984__b _ _ , , 284 , 695 white , r.e .
2003 , , 148 , 487 white , r.e . ,
allen , c.l .
, forrester , w.b . , gonnella , a.m. , & young , k.l .
2001 , , 132 , 253 white , r.e .
, & bally , j. 1993 , , 409 , 234 younan , k.f . , & dufton , p.l .
1984 , , 209 , 123 zsarg , j. , & federman , s.r .
2003 , , 589 , 319 | we present high - resolution , high signal to noise absorption - line observations of cn , , , ch@xmath0 , and ch along twenty lines of sight toward members of the pleiades .
the acquired data enable the most detailed study to date of the interaction between cluster stars and the surrounding interstellar gas .
total equivalent widths are consistent with previous investigations except where weaker features are detected owing to our greater sensitivity .
mean @xmath1-values for the molecular species indicate that toward most of the pleiades ch is associated with the production of ch@xmath0 rather than cn .
an analysis of radial velocities reveals a kinematic distinction between ionized atomic gas and molecular and neutral gas .
molecular components are found with velocities in the local standard of rest of either @xmath2 + 7 km s@xmath3 or @xmath2 + 9.5 km s@xmath3 , with the higher - velocity components associated with the strongest absorption .
atomic gas traced by shows a strong central component at @xmath4 @xmath2 + 7 km s@xmath3 exhibiting velocity gradients indicative of cloud - cluster interactions .
gas density estimates derived from measured ch / ch@xmath0 column density ratios show good agreement with those inferred from h@xmath5 rotational populations , yielding typical values of @xmath6 @xmath7 .
our models do not include the important time - dependent effects on ch@xmath0 formation which may ultimately be needed to extract physical conditions in these clouds . |
liquid chromatography combined with electrospray ionisation mass spectrometry ( esi - ms ) is one of the most frequently used approaches for conducting metabolomics experiments @xcite .
collision - induced dissociation ( cid ) is usually employed within this procedure , intentionally fragmenting molecules into smaller parts to examine their structure .
this is called ms / ms or tandem mass spectrometry . a significant bottleneck in such experiments is the interpretation of the resulting spectra to identify metabolites .
widely used methods for putative metabolite identification @xcite , using mass spectrometry , compare a collected ms or ms / ms spectrum for an unknown compound against a database containing reference ms or ms / ms spectra @xcite .
unfortunately , current reference databases are still fairly limited , especially in the case of esi - ms / ms . at the time of writing , the public human metabolome database @xcite contains esi - ms / ms data for around 800 compounds , which represents only a small fraction of the 40,468 known human metabolites it lists .
the publicly available metlin database @xcite provides esi - ms / ms spectra for 11,209 of the 75,000 endogenous and exogenous metabolites it contains , although more than half of those spectra are for enumerated tripeptides .
the public repository massbank @xcite contains a more diverse dataset of 31,000 spectra collected on a variety of different instruments , including esi - ms / ms spectra for approximately 2000 unique compounds .
however , set against the more than 19 million chemical structures in the pubchem compound database @xcite , an estimated 200,000 plant metabolites @xcite , or even the 32,801 manually annotated entries in the database of chemical entities of biological interest ( chebi ) @xcite , we see that ms / ms coverage still falls far short of the vast number of known metabolites and molecules of interest .
consequently , there is substantial interest in finding alternative means for identifying metabolites for which no reference spectra are available @xcite . for these cases , one approach to metabolite identification involves first predicting the ms or ms / ms spectrum for each candidate compound from its chemical structure @xcite .
the interpreter then uses these predicted spectra in place of reference spectra , and labels the target spectrum as the metabolite whose predicted spectrum is the closest match , according to some similarity criteria . a wide range of similarity criteria
have been proposed , from weighted counts of the number of matching peaks @xcite , to more complex probability based measures @xcite .
the upshot of this predictive approach is that only a list of candidate molecules is needed , rather than a complete database of reference spectra .
however , the restriction to a list of candidate molecules means that this approach still falls short of _ de novo _ identification of unknown unknowns @xcite , -i.e .
we can not identify molecules not in the list .
the concept of computer - based ms prediction has been around since the dendral project in the 1960 s , when investigators attempted to predict electron ionization ( ei ) mass spectra using early machine learning methods @xcite .
more recent approaches to this problem have generally taken one of two forms : rule - based or combinatorial .
commercial packages , such as mass frontier ( thermo scientific , www.thermoscientific.com ) , and ms fragmenter ( acd labs , www.acdlabs.com ) , are rule - based , using thousands of manually curated rules to predict fragmentations .
primarily developed for ei fragmentation , these packages have been extended for use with esi .
this current work does not compare against these methods empirically , however in at least one study they have been found to have been out - performed by metfrag @xcite , to which we do compare
. @xcite another knowledge - based approach , called massimo , combines chemical knowledge with data ; using logistic regression to predict fragmentation probabilities for a particular class of ei fragmentations @xcite .
the other class of algorithms applies a combinatorial fragmentation procedure , enumerating all possible fragments of the original structure by systematically breaking bonds @xcite .
first proposed by @xcite , this method has been incorporated into the freely available programs fid @xcite and metfrag @xcite .
both identify the given spectrum with the metabolite that has the most closely matching peaks via such a combinatorial fragmentation .
these programs also employ several heuristics in their scoring protocols to emphasise the importance of more probable fragmentations .
fid uses an approximate measure of the dissociation energy of the broken bond , combined with a measure of the energy of the product ion .
metfrag incorporates a similar measure of bond energy combined with a bonus if the neutral loss formed is one of a common subset .
an alternative method , fingerid @xcite , takes advantage of the increasing number of available ms / ms spectra , by applying machine learning methods to this task .
this program uses support vector machines ( svms ) to predict a chemical fingerprint directly from an ms / ms spectrum , and then searches for the metabolite that most closely matches that predicted fingerprint . @xcite .
the main problem with the current combinatorial methods is that , while they have very good recall , explaining most if not all peaks in each spectrum , they also have poor precision , predicting many more peaks than are actually observed .
metfrag and fid attempt to address this problem by adding the heuristics described above . in our work ,
we investigate an alternative machine learning approach that aims to improve the precision of such combinatorial methods .
we propose a method for learning a generative model of the cid fragmentation process from data .
this model estimates the likelihood of any given fragmentation event occurring , thereby predicting those peaks that are most likely to be observed .
we hypothesise that increasing the precision of the predicted spectrum in this way will improve our system s ability to accurately identify metabolites .
@xcite section [ sec : cfm ] provides details of our proposed model and the training method .
section [ sec : results ] then reports the experimental results .
we will assume the reader knows the foundations of esi ms / ms ; for an introduction to this process , see @xcite .
[ [ sec : cfm ] ] this section presents our model for the esi - ms / ms cid fragmentation process , which we call competitive fragmentation modeling ( cfm ) , and a method for deriving parameters for this model from existing ms / ms data .
section [ sec : se ] describes the simplest form of this method ; single energy competitive fragmentation modeling ( se - cfm ) .
section [ combinedenergymodel ] then presents an extension of this method , combined energy competitive fragmentation modeling ( ce - cfm ) , which aims to make better use of cid ms / ms spectra measured at different energy levels for the same compound .
windows executables and cross - platform source code and the trained models are freely available at http://sourceforge.net / projects / cfm - id/. a web server interface is also provided at http://cfmid.wishartlab.com .
section [ sec : results ] in single energy cfm ( se - cfm ) , we model esi - ms / ms fragmentation as a stochastic , homogeneous , markov process @xcite involving state transitions between charged fragments , as depicted in figure [ fig : model](a ) . [
fig : model ] more formally , the process is described by a fixed length sequence of discrete , random fragment states @xmath0 , where each @xmath1 takes a value from the state space @xmath2 , the set of all possible fragments ; this state space will be further described in section [ statespace ] .
a transition model defines the probabilities that each fragment leads to another at one step in the process ; see section [ transitionmodel ] .
an observation model maps the penultimate node @xmath3 to a peak @xmath4 , which takes on a value in @xmath5 that represents the m / z value of the peak to which the final fragment will contribute ; see section [ observationmodel ] .
se - cfm is a latent variable model in which the only observed variables are the initial molecule @xmath6 and the output peak @xmath4 ; the fragments themselves are never directly observed .
each output @xmath4 adds only a small contribution to a single peak in the mass spectrum . in order to predict a complete mass spectrum
, we can run the model forward multiple times to compute the marginal distribution of @xmath4 .
we make the following assumptions about the cid fragmentation process .
further details for the motivations of each are provided below , but these generally involve a trade - off between accurately modeling the process and keeping the model computationally tractable . 1 .
all input molecules have a single positive charge and exist in their most common isotopic form
. 2 . in a collision
, each molecule will break into two fragments .
no mass or charge is lost .
one of the two fragments must have a single positive charge and the other must be neutral .
combined , the two must contain all the components of the original charged molecule , i.e. all the atoms and electrons .
4 . no further sigma bonds can be removed or added during a break , except those connecting hydrogens i.e .
the edges in the molecular graph must remain the same .
rearrangement of pi bonds is allowed and hydrogen atoms may move anywhere in the two resulting fragments , on the condition that both fragments satisfy all valence rules , and standard bond limitations are met e.g .
no bond orders higher than triple .
the even electron rule is always satisfied i.e .
no radicals .
assumption 1 is reasonable as we assume that the first phase of ms / ms successfully restricts the mass range of interest to include only the [ m+h]@xmath7 precursor ion containing the most abundant isotopes .
since this ion has only a single positive charge , we can safely assume that no multiply - charged ions will be formed in the subsequent ms2 phase .
ensuring that valid [ m+h]@xmath7 precursor ions are selected in ms1 is beyond the scope of this work ; see @xcite for a summary of ms1 data processing methods .
assumptions 2 , 4 and 6 do not necessarily hold in real - world spectra @xcite .
however including them substantially reduces the branching factor of the fragment enumeration , making the computations feasible . since these assumptions do appear to hold in the vast majority of cases
, we expect that including them should have minimal negative impact on the experimental results .
note that most 3-way fragmentations can be modeled by two sequential , 2-way fragmentations , so including assumption 2 should not impact our ability to model most fragmentation events .
assumption 5 allows for mclafferty rearrangement and other known fragmentation mechanisms @xcite .
our method for enumerating fragments is similar in principle to the combinatorial approach used in metfrag and fid @xcite , with some additional checks to enforce the above assumptions .
we systematically break all non - ring bonds in the molecule ( excluding those connecting to hydrogens ) and all pairs of bonds within each ring .
we do this one break at a time , enumerating a subset of fragments with all possible masses that may form after each break , .
this subset is found by determining the number of additional electrons that can be allocated to either side of the break using integer linear programming e.g . breaking the middle bond in ccc[ch4 + ]
( smiles format ) gives fragments c=[ch3 + ] ( mass=29.04da , loss cc ) and c[ch4 + ] ( mass=31.05da , loss c = c ) , the fragmentation procedure is applied recursively on all the produced fragments , to a maximum depth .
the result is a directed acyclic graph ( dag ) containing all possible charged fragments that may be generated from that molecule .
an abstract example of such a fragmentation graph is provided in figure [ fig : fragmentationgraph ] .
note that for each break , one of the two produced fragments will have no charge .
since it is not possible for a mass spectrometer to detect neutral molecules , we do not explicitly include the neutral fragments in the resulting graph , nor do we recur on their possible breaks .
however neutral loss information may be included on the edges of the graph , indicating how a particular charged fragment was determined .
@xcite our parametrized transition model assigns a conditional probability to each fragment given the previous fragment in the sequence @xmath6,@xmath8, ,@xmath3 . in the case where @xmath9 has @xmath10 as a possible child fragment in a fragmentation graph , our model assigns a positive probability to the transition from @xmath11 to @xmath12 .
furthermore , self - transitions are always allowed , i.e. the probability of transitioning from @xmath11 to @xmath13 is always positive ( for the same @xmath9 ) .
we assign 0 probability to all other transitions , i.e. those that are not self - transitions , and that do not exist within any fragmentation graph .
although the set of possible charged fragments @xmath14 is large , the subset of child fragments originating from any particular fragment is relatively small .
for example , the requirement that a feasible child fragment must contain a subset of the atoms in the parent fragment rules out many possibilities .
consequently most transitions will be assigned a probability of 0 .
note that the assigned probabilities of all transitions originating at a particular fragment , including the self - transition , must sum to one .
we now discuss how we parametrize our transition model .
a natural parametrization would be to use a transition matrix containing a separate parameter for every possible fragmentation @xmath15 .
unfortunately , we lack sufficient data to learn parameters for every individual fragmentation in this manner . instead , we look for methods that can generalize by exploiting the tendency of similar molecules to break in similar ways . we introduce the notion of _ break tendency _ , which we represent by a value @xmath16 for each possible fragmentation @xmath15 that models how likely a particular break is to occur .
those fragmentations that are more likely to occur are assigned a higher break tendency value , and those that are less likely are given lower values .
we then employ a softmax function to map the break tendencies for all breaks involving a particular parent fragment to probabilities , as defined in equation [ rhoequation ] below .
this has the effect of capturing the competition that occurs between different possible breaks within the same molecule .
for example , consider the two fragmentations in figure [ fig : twobreaks ] . here ,
although both fragmentations involve an h@xmath17o neutral loss , in the left - hand case , the h@xmath17o loss must compete with the loss of an ammonia group , whereas in the right hand case , it does not .
hence our model might assign an equal break tendency to both cases , but this would still result in a lower probability of fragmentation in the former case , due to the competing ammonia . two similar breaks , both resulting in an h@xmath17o neutral loss . the right case should be assigned a higher probability , as in the left case , the nh@xmath18 is also likely to break away , reducing the probability of the h@xmath17o loss . ]
we model the probability of a particular break @xmath19 occurring as a function of its break tendency value @xmath20 and that of all other competing breaks from the same parent , as follows : @xmath21 where the sums iterate over all @xmath22 for which @xmath23 is possible .
since the break tendency is a relative measure , it makes sense to tie it to some reference point . for the purposes of this model ,
we have assigned the break tendency for a self - transition ( i.e. no break occuring ) to @xmath24 , which gives @xmath25 as shown in ( [ rhoequation ] ) .
[ [ incorporating - chemical - features ] ] incorporating chemical features + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we need to compute @xmath20 for @xmath26 . to do this
we first define a binary feature vector @xmath27 to describe the characteristics of a given break @xmath15 .
such features might include the presence of a particular atom adjacent to the broken bond , or the formation of a specific neutral loss molecule e.g .
see section [ sec : chemicalfeatures ] .
we then use these features to assign a break tendency value using a linear function parameterized by a vector of weights @xmath28 i.e .
@xmath29 . this
can then be substituted into ( [ rhoequation ] ) to generate the probability of transition @xmath19 .
the first feature of @xmath27 is a bias term , set to 1 for all breaks .
note that the vector @xmath30 constitutes the parameters of the cfm model that we will be learning .
we model the conditional probability of @xmath4 using a narrow gaussian distribution centred around the mass ) and hence can just use the mass here .
] of @xmath3 , i.e. @xmath31 .
the value for @xmath32 can be set according to the mass accuracy of the mass spectrometer used .
so , we define this observation function to be the following @xmath33 our investigation ( see supplementary data ) of the mass error of the precursor ions in the metlin metabolite data used in section [ sec : results ] found that the distribution of mass errors had a mean offset of approximately 1 ppm , and a narrower shape than a gaussian distribution .
however , in order to model a more general mass error , not specific to a particular instrument or set of empirical data , we think the gaussian distribution is a reasonable approach .
our system estimates the values for the parameters @xmath30 of the proposed model by applying a training procedure to a set of molecules @xmath34 , for which we have both the chemical structure and a measured ms / ms spectrum . for the purposes of this work ,
we assume we have a measured low , medium and high energy cid ms / ms spectrum for each molecule , which we denote @xmath35 ( @xmath36 , @xmath37 , @xmath38)@xmath39 each spectrum is further defined to be a set of peaks , where each peak is a pair @xmath40 , composed of a mass @xmath41 and a height ( or intensity ) @xmath42 \subset \mathbb{r}$ ] .
note that each spectrum is normalized , such that the peak heights sum to 100 . for this single energy version of the model ,
we derive parameters for a completely separate model for each of the three energy levels , using data from that level only .
note that if we had data for only one energy level , we could use this method to train a model using just that energy .
however section [ combinedenergymodel ] will extend this model to combine the three energy spectra for use in a single model . until then , we will use @xmath43 to denote whichever of @xmath36 , @xmath37 or @xmath38 we are currently considering .
[ [ maximum - likelihood ] ] maximum likelihood + + + + + + + + + + + + + + + + + + we use a maximum likelihood approach for parameter estimation .
the likelihood of the data @xmath44 , given the parameters @xmath30 , and incorporating the previously defined transition function @xmath45 and observation function @xmath46 , is given by @xmath47 where @xmath48 denotes the children of @xmath9 in all fragmentation graphs containing it , and @xmath49 .
however we are unable to maximize this function in closed form .
instead we use the iterative expectation maximization @xcite technique .
[ [ expectation - maximization - em ] ] expectation maximization ( em ) + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in the e - step , the expected log likelihood expression is given by @xmath50 where @xmath51 denotes the values for @xmath30 on the @xmath52-th iteration .
substituting ( [ rhoequation ] ) and ( [ obsequation ] ) into the above and re - arranging in terms of all possible fragment pairs gives @xmath53 where @xmath54 @xmath55 and @xmath56 in the m - step , we look for the @xmath51 that maximizes the above expression of @xmath57 . noting that @xmath58 is independent of @xmath51 and denoting the @xmath59th component of @xmath30 as @xmath60 , @xmath61\phi_{i , j}^{l } - \hspace{-8pt}\sum\limits_{k\in c(f_{i})}\hspace{-8pt}\phi_{i , k}^{l}\rho(f_{i},f_{k};w)\big)\ ] ] where @xmath62 denotes the @xmath59th component of the feature vector @xmath63 and @xmath64 $ ] is the indicator function .
this does not permit a simple closed - form solution for @xmath30 .
however @xmath65 is concave in @xmath51 , so settings for @xmath51 can be found using gradient ascent .
values for the joint probabilities in the @xmath66 terms can be computed efficiently using the junction tree algorithm @xcite .
we also add an @xmath67 regularizer on the values of @xmath30 to @xmath57 ( excluding the bias term ) .
this has the effect of discouraging overfitting by encouraging the parameters to remain close to zero .
ms / ms spectra are often collected at multiple collision energies for the same molecule .
increasing the collision energy usually causes more fragmentation events to occur .
this means that fragments appearing in the medium and high energy spectra are almost always descendants of those that appear in the low and medium energy spectra , respectively .
so the existence of a peak in the medium energy spectrum may help to differentiate between explanations for a related peak in the low or high energy spectra .
for this reason , we also assessed an additional model , combined energy cfm ( ce - cfm ) , which extends the se - cfm concept by combining information from multiple energies as shown in fig .
[ model ] ( b ) .
plow , pmed and phigh each represent a peak from the low , medium and high energy spectrum respectively .
the fragment states , transition rules and the observation model are all the same here as for se - cfm .
the main difference now is that the homogeneity assumption is relaxed so that separate transition likelihoods can be learned for each energy block i.e .
, @xmath6 to @xmath68 , @xmath68 to @xmath69 and @xmath69 to @xmath70 , where @xmath71 , @xmath72 and @xmath73 denote the fragmentation depths of the low , medium and high energy spectra respectively .
this results in separate parameter values for each energy , denoted respectively as @xmath74 , @xmath75 and @xmath76 .
the complete parameter set for this model thus becomes @xmath77 .
we can again use a maximum likelihood approach to parameter estimation based on the em algorithm .
this approach deviates from the se - cfm method only as follows : * for each energy level , ( [ gradequation ] ) is computed separately , restricting the @xmath78 terms to relevant parts of the model e.g .
@xmath79 would sum from @xmath80 to @xmath72 when computing the gradients for @xmath75 , and from @xmath81 to @xmath73 when computing gradients for @xmath76 . *
the computation of the @xmath82 terms combines evidence from the full set of three spectra @xmath83 . in se - cfm
, we apply one spectrum at a time , effectively sampling from a distribution over the peaks from each observed spectra . in this extended model
we can not do this because we do not have a full joint distribution over the peaks , but rather we only have marginal distributions corresponding to each spectrum .
the standard inference algorithms e.g .
the junction tree algorithm , do not allow us to deal with observations that are marginal distributions rather than single values .
instead we use the iterative proportional fitting procedure ( ipfp ) @xcite , with minor modifications to better handle cases where the spectra are inconsistent ( not simultaneously achievable under any joint distribution ) .
these modifications reassign the target spectra to be the average of those encountered when the algorithm oscillates in such circumstances .
in this section we present results using the above described se - cfm ( @xmath84=2 ) and ce - cfm ( @xmath71=2 , @xmath72=4 , @xmath73=6 ) methods , on a spectrum prediction task , and then in a metabolite identification task .
we used the metlin database @xcite , separated into two sets ( see description below ) each containing positive mode , esi - ms / ms spectra from a 6510 q - tof ( agilent technologies ) mass spectrometer , measured at three different collision energies : 10v , 20v and 40v , which we assign to be low , medium and high energy respectively .
each set was randomly divided into 10 groups for use within a 10-fold cross validation framework . 1 .
* tripeptides * : the metlin database contains data for over 4000 enumerated tripeptides .
we randomly selected 2000 of these molecules , then omitted 15 that had four or more rings due to computational resource concerns , leaving 1985 remaining in the set .
fragmentation patterns in peptides are reasonably well understood @xcite , leading to effective algorithms for identifying peptides from their esi ms / ms data e.g .
@xcite . however , we think that the size of this dataset , and the fact that it contains so many similar yet different molecules , make it an interesting test case for our algorithms .
* metlin metabolites * : we use a set of 1491 non - peptide metabolites from the metlin database .
these are a more diverse set covering a much wider range of molecules .
an initial set of 1500 were selected randomly .
nine were then excluded because they were so much larger than the other molecules ( over 1000 da ) , such that their fragmentation graphs could not be computed in a reasonable amount of time .
we also used an additional small validation set , selected because they were measured on a similar mass spectrometer , an agilent 6520 q - tof , but in a different laboratory .
these were taken from the massbank database @xcite .
all testing with this set used a model trained for the first cross - fold set of the metlin metabolite data ( @xmath85 of the data ) . 1 .
* massbank metabolites * : this set contains 192 metabolites taken from the washington state university submission to the massbank database .
all molecules from this submission were included that had ms2 spectra with collision energies 10v , 20v and 40v , in order to provide a good match with the metlin data .
files containing test molecule lists and assigned cross validation groups are provided as supplementary data .
the chemical features used in these experiments were as follows .
note that the terms _ ion root atom _ and _ neutral loss ( nl ) root atom _ refer to the atoms connected to the broken bond(s ) on the ion and neutral loss sides respectively cf . , fig .
[ fig : breaklabel ] . * _ break atom pair _ : indicators for the pair of ion and neutral loss root atoms , each from \{c , n , o , p , s , other } , included separately for those in a non - ring break vs those in a ring break e.g .
[ fig : breaklabel]a ) : would be non - ring c - c .
( 72 features ) * _ ion and nl root paths _ indicators for all paths of length 2 and 3 starting at the respective root atoms and stepping away from the break .
each is an ordered double or triple from \{c , n , o , p , s , other } , taken separately for rings and non - rings .
two more features indicate no paths of length 2 and 3 respectively e.g .
[ fig : breaklabel]a ) : the ion root paths are c - o , c - n and c - n - c .
( 2020 features ) . * _ gasteiger charges _ : indicators for the quantised pair of gasteiger charges @xcite for the ion and nl root atoms in the original unbroken molecule .
( 288 features ) * _ hydrogen movement _ : indicator for how many hydrogens switched sides of the break and in which direction i.e .
ion to nl ( - ) or nl to ion(+ ) \{0,@xmath86,@xmath87,@xmath88,@xmath89,other}. ( 10 features ) * _ ring features _ : properties of a broken ring .
aromatic or not ?
multiple ring system ?
size \{3,4,5,6 , other } ?
distance between the broken bonds \{1,2,3,4 + } ?
[ fig : breaklabel]b ) is a break of a single aromatic ring of size 6 at distance 3 .
( 12 features ) . of these 2402 features ,
few take non - zero values for any given break .
many are never encountered in our data set , in which case their corresponding parameters are set immediately to 0 .
we also append _
quadratic features _
, containing all 2,881,200 pair - wise combinations of the above features , excluding the additional bias term .
again , most are never encountered , so their parameters are set to 0 . for each cross validation fold , and the massbank validation set , a model ( trained as above ) , was used to predict a low , medium and high energy spectra for each molecule in the test set .
the resulting marginal distributions for the peak variables are a mixture of gaussian distributions .
we take the means and weights of these gaussians as our peak mass and intensity values .
since all fragments in the fragmentation graph of a molecule have non - zero probabilities in the marginal distribution , it is necessary to place a cut - off on the intensity values to select only the most likely peaks . here
, we use a post - processing step that removes peaks with low probability , keeping as many of the highest peaks as required to form at least 80% of the total intensity sum .
we also set limits on the number of selected peaks to be at least 5 and at most 30 .
this ensures that more peaks are included than just the precursor ion , and also prevents spectra occurring that have large numbers of very small peaks .
these values were selected arbitrarily , but post - analysis suggests that they are reasonable ( see supplementary data ) .
when matching peaks we use a mass tolerance set to the larger of 10 ppm and 0.01 da ( depending on the peak mass ) , and set the observation parameter @xmath32 to be one third of this value .
[ [ metrics ] ] metrics + + + + + + + we consider a peak in the predicted ms / ms spectrum @xmath90 to match a peak in the measured ms / ms spectrum @xmath91 if their masses are within the mass tolerance above .
we use the following metrics : 1 .
* weighted recall * : the percentage of the total peak intensity in the measured spectrum with a matching peak in the predicted spectrum : @xmath92 \hspace{2pt}\div \hspace{-10pt}\sum\limits_{(m , h)\in s_{m}}\hspace{-12pt}h$ ] .
2 . * weighted precision * : the percentage of the total peak intensity in the predicted spectrum with a matching peak in the measured spectrum : @xmath93 \hspace{2pt}\div \hspace{-10pt}\sum\limits_{(m , h)\in s_{p}}\hspace{-12pt}h$ ] .
* recall * : the percentage of peaks in the measured spectrum that have a matching peak in the predicted spectrum : @xmath94 .
* precision * : the percentage of peaks in the predicted spectrum that have a matching peak in the measured spectrum : @xmath95 .
* jaccard score * : @xmath96 .
the intensity weighted metrics were included because the unweighted precision and recall values can be misleading in the presence of low - level noise e.g . when there are many small peaks in the measured spectrum .
the weighted metrics place a greater importance on matching higher intensity peaks , and therefore give a better indication of how much of a spectrum has been matched .
however , these weighted metrics can also be susceptible to an over - emphasis of just one or two peaks , and in particular of the peak corresponding to the precursor ion .
consequently , we think it is informative to consider both weighted and non - weighted metrics for recall and precision .
[ [ models - for - comparison ] ] models for comparison + + + + + + + + + + + + + + + + + + + + + : the pre - existing methods , e.g .
metfrag , fingerid do not output a predicted spectrum , but skip directly to metabolite identification .
so , instead we compare against : * * full enumeration * : this model considers the predicted spectrum to be one that enumerates all possible fragments in the molecule s fragmentation tree with uniform intensity values . * * heuristic * ( tripeptides only ) : this model enumerates known peptide fragmentations as described by @xcite , including @xmath97 , @xmath98 , @xmath99 , @xmath100 , @xmath101 , @xmath102 and immonium ions . [ [ results ] ] results + + + + + + + : the results are presented in figure [ fig : metricspeptides ] . for all three data sets , se - cfm and ce - cfm obtain several orders of magnitude better precision and jaccard scores than the full enumerations of possible peaks .
there is a corresponding loss of recall .
however , if we take into account the intensity of the measured peaks , by considering the weighted recall scores , we see that our methods perform well on the more important , higher intensity peaks .
more than 75% of the total peak intensity in the tripeptide spectra , and approximately 60% of the total peak intensity in the metabolite spectra , were predicted .
the results presented in figure [ fig : metricspeptides ] show scores averaged across the three energy levels for each molecule . if we consider the results for the energy levels separately ( see supplementary data ) , we find that the low and medium energy results are much better for all methods we assessed . for example , in the case of the low energy spectra , the weighted recall scores for se - cfm are 78% , 73% and 81% for the tripeptide , metlin metabolite and massbank metabolite data sets respectively , as compared to 73% , 29% and 37% respectively for the high energy spectra .
the poorer high energy spectra results may be due to increased noise and a lower predictability of events at the higher collision energies .
another possible explanation is that the even - electron rule and other assumptions listed in section [ statespace ] may be less reliable when there is more energy in the system . in the case of the tripeptide data ,
our methods achieve higher recall scores and similar rates of precision to that of the heuristic model of known fragmentation mechanisms , resulting in improved jaccard scores .
since peptide fragmentation mechanisms are fairly well understood , this result is not intended to suggest that our method should be used in place of current peptide fragmentation programs , but rather to demonstrate that se - cfm and ce - cfm are able to extract fragmentation patterns from data to a similar extent to human experts , given a sufficiently large and consistent data set . like our methods ,
the heuristic models also perform better for the lower energy levels , with a weighted recall score of 66% for the low energy , as compared to only 24% for the high energy
. unsurprisingly , being a smaller and more diverse data set , the metlin metabolite results are poorer than those of the tripeptides .
however the weighted recall for both our methods is still above 60% and the precision and jaccard scores are much higher than for the full enumeration , suggesting that the cfm model is still able to capture some of the common fragmentation trends .
the weighted recall and precision results for the massbank metabolites are fairly comparable to those of the metlin metabolites .
there is a small loss in the non - weighted recall , however this is probably due to a higher incidence of low - level noise in the massbank data .
this results in a small loss in the average jaccard score .
however these results demonstrate that the fragmentation trends learned still apply to a significant degree on data collected at a different time in a different laboratory .
since this is the first method , to the authors knowledge , capable of predicting intensity values as well as m / z values , we also investigated the accuracy of cfm s predicted intensity values .
we found that the pearson correlation coefficients for matched pairs of predicted and measured peaks , were 0.7 , 0.6 and 0.45 for the low , medium and high spectra respectively ( se - cfm and ce - cfm results were not significantly different ) .
this indicates a positive , though imperfect correlation .
full results and scatter plots are contained in the supplementary data . here
we apply our cfm ms / ms spectrum predictions to a metabolite identification task .
for each molecule , we produce two candidate sets via queries to two public databases of chemical entities : 1 .
we query the pubchem compound database @xcite for all molecules within 5 ppm of the known molecule mass .
this simulates the case where little is known about the candidate compound , but the parent ion mass is known with high accuracy .
2 . we query kegg ( kyoto encyclopedia of genes and genomes ) @xcite for all the molecules within 0.5 da of the known molecular mass .
this simulates the case where the molecule is thought to be a naturally occurring metabolite , but there is more uncertainty in the target mass range . to conduct this assessment , duplicate candidates were filtered out i.e .
those with the same chemical structure , including those that only differ in their stereochemistry .
charged molecules and ionic compounds were also removed since the program assumes single fragment , neutral candidates ( to which it will add a proton ) . after filtering , the median number of candidates returned from pubchem was 911 for the tripeptides and 1025 for the metabolites .
note that 9 tripeptides and 57 of the metlin metabolites were excluded from this testing because no matching entry was found in pubchem for these molecules .
the kegg queries were only carried out for the metabolite data .
the median number of candidates returned was 22 , however no matching entry was found in kegg for 833 of the metlin metabolites and 111 of the massbank metabolites .
whenever a matching entry could be found , we ranked the candidates according to how well their predicted low , medium and high spectra matched the measured spectra of the test molecule .
the ranking score we used was the jaccard score described in section [ sec : spectrumprediction ] .
we compared the ranking performance of our se - cfm and ce - cfm methods against those of metfrag @xcite and fingerid @xcite .
we used the same candidate lists for all programs .
for candidate molecules with equal scores , we had each program break ties in a uniformly random manner .
this was in contrast to the original metfrag code , which used the most pessimistic ranking ; we did not use that approach as it seemed unnecessarily pessimistic .
we set the mass tolerances used by metfrag when matching peaks to the same as those used in our method ( maximum of 0.01da and 10ppm ) .
metfrag and fingerid only accept one spectrum , so to input the three spectra we first merged them as described by @xcite : we took the union of all peaks , and then merge together any peaks within 10 ppm or 0.01 da of one another , retaining the average mass and the maximum intensity of the two .
in fingerid we used the linear high resolution mass kernel including both peaks and neutral losses , and trained using the same cross - fold sets as for our own method .
overall , we attempted to assess cfm , metfrag and fingerid as fairly as possible , using identical constraints , identical databases and near - identical data input .
the results are shown in figure [ fig : rankingresults ] .
as seen in this figure , our cfm method achieved substantially better rankings than both the existing methods on all three data sets , for both the pubchem and kegg queries .
when querying against kegg , our methods found the correct metabolite as the top - scoring candidate in over 70% of cases for both metabolite sets and almost always @xmath103 ranked the correct candidate in the top 5 . in comparison
, metfrag ranked the correct metabolite first in approximately 50% of cases for both metabolite sets , and in the top 5 in 89% .
fingerid ranked the correct metabolite first in less than 15% of cases .
for pubchem , our methods performed well on the tripeptide data , identifying the correct metabolite as the top - scoring candidate in more than 50% of cases and ranking the correct candidate in the top 10 for more than 98% of cases .
this is again convincingly better than both metfrag and fingerid , which rank the correct candidate first in less than 35% and 2% of cases respectively . for the metabolite data , ce - cfm and se - cfm were able to identify the correct metabolite in only 12% and 10% of cases respectively , however given that this is from a list of approximately one thousand candidates , this performance is still not bad .
once again , it is substantially better than metfrag and fingerid , which correctly identified less than 6% and 1% of cases respectively .
our methods rank the correct candidate in the top 10 in more than 40% of cases on both data sets , as compared to metfrag s performance of 31% on the metlin metabolites and 21% on the massbank metabolites .
additionally , the top - ranked compound was found to have the correct molecular formula in more than 88% of cases for se - cfm and 90% of cases for ce - cfm , suggesting that both methods mainly fail to distinguish between isomers . while the performance of all three methods ( cfm , metfrag and fingerid ) is not particularly impressive for the pubchem data sets ( i.e. @xmath10412% correct ) we would argue that the pubchem database is generally a poor database choice for anyone wishing to do ms / ms metabolomic studies . with only 1% of its molecules having a biological or natural product origin , one is already dealing with a rather significant challenge of how to eliminate a 100:1 excess of false positives .
so we would regard the results from the pubchem assessment as a `` worst - case '' scenario and the results from the kegg assessment as a more typical metabolomics scenario . the results for ce - cfm showed minimal difference when compared to those of se - cfm , casting doubt on whether the additional complexity of ce - cfm is justified .
however we think this idea is still interesting as a means for integrating information across energy levels and may yet prove more useful in future work .
( see section [ sec : spectrumprediction ] ) , http://cfmid.wishartlab.com we encourage readers to make use of our executables and source code , made available at http://sourceforge.net / projects / cfm - id/.
we have proposed a model for the esi - ms / ms fragmentation process and a method for training this model from data .
the performance has been benchmarked in cross validation testing on a large molecule set , and further validated using an additional dataset from another laboratory .
head - to - head comparisons using multiple data sets under multiple conditions show that the cfm method significantly outperforms existing state - of - the - art methods , and has attained a level that could be useful to experimentalists performing metabolomics studies .
bolton e , wang y , thiessen p , bryant s ( 2008 ) pubchem : integrated platform of small molecules and biological activities . in : chapeter 12 in annual reports in computational chemistry ,
vol 4 , american chemical society , washington dc galezowska a , harrison mw , herniman jm , skylaris ck , langley gj ( 2013 ) a predictive science approach to aid understanding of electrospray ionisation tandem mass spectrometric fragmentation pathways of small molecules using density functional calculations .
rapid communications in mass spectrometry : rcm 27(9):964970 hill
aw , mortishire - smith rj ( 2005 ) automated assignment of high - resolution collisionally activated dissociation mass spectra using a systematic bond disconnection approach .
rapid communications in mass spectrometry 19(21):31113118 kangas lj , metz to , isaac g , schrom bt , ginovska - pangovska b , wang l , tan l , lewis rr , miller jh ( 2012 ) in silico identification software ( isis ) : a machine learning approach to tandem mass spectral identification of lipids .
bioinformatics 28(13):170513 levsen k , schiebel
hm , terlouw jk , _
et al_(2007 ) even - electron ions : a systematic study of the neutral species lost in the dissociation of quasi - molecular ions .
journal of mass spectrometry : jms 42:10241044 oberacher h , pavlic m , libiseller k , _
et al _ ( 2009 ) on the inter - instrument and the inter - laboratory transferability of a tandem mass spectral reference library : 2 .
optimization and characterization of the search algorithm .
journal of mass spectrometry : jms 44(4):494502 | electrospray tandem mass spectrometry ( esi - ms / ms ) is commonly used in high throughput metabolomics .
one of the key obstacles to the effective use of this technology is the difficulty in interpreting measured spectra to accurately and efficiently identify metabolites .
traditional methods for automated metabolite identification compare the target ms or ms / ms spectrum to the spectra in a reference database , ranking candidates based on the closeness of the match .
however the limited coverage of available databases has led to an interest in computational methods for predicting reference ms / ms spectra from chemical structures .
this work proposes a probabilistic generative model for the ms / ms fragmentation process , which we call competitive fragmentation modeling ( cfm ) , and a machine learning approach for learning parameters for this model from ms / ms data .
we show that cfm can be used in both a ms / ms spectrum prediction task ( ie , predicting the mass spectrum from a chemical structure ) , and in a putative metabolite identification task ( ranking possible structures for a target ms / ms spectrum ) . in the ms / ms spectrum prediction task , cfm shows significantly improved performance when compared to a full enumeration of all peaks corresponding to substructures of the molecule . in the metabolite identification task , cfm obtains substantially better rankings for the correct candidate than existing methods ( metfrag and fingerid ) on tripeptide and metabolite data , when querying pubchem or kegg for candidate structures of similar mass .
example.eps gsave newpath 20 20 moveto 20 220 lineto 220 220 lineto 220 20 lineto closepath 2 setlinewidth gsave .4 setgray fill grestore stroke grestore |
cool white dwarfs have recently gained attention due to the discovery of faint ( @xmath9 ) objects in the hubble deep field survey ( see * ? ? ? * and references therein ) which have been interpreted by @xcite as being cool ( @xmath10 k ) and old ( @xmath11 gyr ) hydrogen atmosphere white dwarfs .
hansen has shown that the location of these objects _ blueward _ of the white dwarf cooling sequence in the @xmath12 vs @xmath13 color - magnitude diagram is the result of the extremely strong h@xmath5 molecular absorption features which extend well into the optical regions of their energy distributions . hence ,
even though hydrogen - rich white dwarfs become redder with decreasing effective temperature down to @xmath14 k ( see , e.g. , fig . 6 of * ? ? ?
* ) , even cooler stars actually become bluer .
note that the most recent photometric and spectroscopic analysis of cool white dwarfs in the local galactic disk ( * ? ? ?
* hereafter brl ) has shown no evidence of white dwarfs cooler than 4000 k. in addition , @xcite has also discussed the importance of using detailed stellar atmospheres as boundary conditions for calculating cooling white dwarf models .
for instance , the cooling age of a 0.6 @xmath6 , @xmath15 k white dwarf can differ by @xmath16 gyr depending on whether the object has a hydrogen- or a helium - dominated atmosphere . given this sensitivity ,
it is important to determine precisely the atmospheric composition of cool white dwarfs before any attempt to determine their age is made . to complicate things even more
, white dwarfs are not expected to preserve their hydrogen or helium atmospheric signature throughout their evolution .
indeed , there is both theoretical and observational evidence which indicates that the chemical composition of white dwarfs is likely to change many times during their lifetime due to competing mechanisms ( e.g. , gravitational settling , convective mixing , convective dredge - up , accretion from the interstellar medium , radiative acceleration ) which may affect the composition of the outer layers of these stars as they cool ( see * ? ? ? * for a review on the subject ) .
this is even more important for cool white dwarfs where convective mixing of a thin ( @xmath17 ) hydrogen atmosphere with the more massive convective helium envelope can turn a hydrogen - rich atmosphere into a helium - dominated atmosphere anywhere between 12,000 k and 5000 k. the exact temperature at which convective mixing occurs or whether it will occur at all depends on the thickness of the hydrogen layer ( the thicker the hydrogen layer , the cooler the mixing temperature ) .
furthermore , brl have clearly demonstrated that the chemical evolution of cool white dwarfs is even more complex than previously anticipated , and that several mechanisms some of them still unknown are at play to alter the chemical constituent of their atmospheres as they cool .
consequently , the age of any cool white dwarf is not only strongly dependent on its atmospheric composition , as pointed out by @xcite , _ but also on its chemical evolution history_. for instance , a recently mixed helium - rich atmosphere white dwarf would have an integrated cooling age of its immediate predecessor i.e. that of a hydrogen - rich atmosphere white dwarf .
stellar ages and chemical evolution can thus not be easily disentangled .
this unfortunate consequence stresses the importance of better understanding the chemical evolution of cool white dwarf stars . for several years
now , we have been pursuing our comprehensive investigation of cool white dwarf evolution .
we presented in brl photometric and spectroscopic observations of 110 cool white dwarfs , and analyzed them with model atmosphere calculations appropriate for cool white dwarfs with pure hydrogen and pure helium compositions , as well as with mixed h / he compositions .
our fitting technique consists of comparing the observed energy distributions in the optical @xmath1 and infrared @xmath2 with those predicted from model atmospheres in order to determine the effective temperature of the star .
when available , trigonometric parallax measurements are used to derive the radius of the star , and hence the mass through the mass - radius relation for white dwarfs . finally , high signal - to - noise spectroscopy at h@xmath3is used to determine , or at least constrain , the chemical composition of their atmospheres .
the main results of the brl analysis are summarized here together with our current understanding of the physical models which have been proposed to account for the observations .
most cool white dwarfs have energy distributions which are consistent with either pure hydrogen or pure helium atmospheric compositions , with little evidence for objects with mixed h / he compositions . in general , the energy distributions of da ( hydrogen - line ) stars are consistent with pure hydrogen atmospheres , while those of non - da stars are consistent with pure helium atmospheres .
exceptions include a few extremely cool non - da stars with hydrogen - dominated atmospheres which are simply too cool to exhibit any hydrogen absorption features , the so - called c@xmath5h stars whose peculiar energy distributions and strong molecular absorption bands have been interpreted as the result of mixed h / he / c compositions with @xmath18 @xcite , and a few other objects which are most likely unresolved binaries ( see brl and references therein ) . note that hydrogen lines are observed in da stars as cool as @xmath19 k , that is , almost at the end of the observable white dwarf cooling sequence of the local galactic disk .
brl have also shown that the temperature distribution of da and non - da stars is not homogeneous .
in particular , even though non - da stars are abundant above @xmath20 k , most white dwarfs below this temperature , down to @xmath15 k , are of the da type ( the so - called `` non - da gap '' ) . moreover , brl have identified a group of objects which lie close to , and above , @xmath20 k whose energy distributions are better reproduced with pure hydrogen models , even though their optical spectra are completely featureless ; hydrogen atmosphere white dwarfs in this particular range of temperature should exhibit strong hydrogen lines .
brl proposed an exotic model to account for both the strange nature of these objects and the existence of the non - da gap .
a warmer , pure helium atmosphere white dwarf slowly accretes hydrogen from the interstellar medium as it cools .
most atomic levels of the hydrogen accreted are quenched as a result of the high atmospheric pressures which characterize cool , helium - rich atmospheres .
in particular , the bound level of @xmath21 is destroyed , and the corresponding bound - free @xmath21 opacity the most important opacity source in hydrogen - rich atmospheres is significantly reduced .
consequently , the accretion of hydrogen onto a helium - dominated atmosphere would have little effect on the opacity and thus on the atmospheric pressure .
brl proposed that large amounts of hydrogen could be _ hidden _ through this simple accretion / quenching mechanism .
they also suggested that at some point , the free - free @xmath21 opacity would become large enough for the atmospheric pressure to drop below the threshold value at which the @xmath21 ion starts to recombine .
the star then rapidly turns into a normal hydrogen - rich white dwarf .
brl suggested that the group of peculiar white dwarfs discussed above could represent such objects with a mixture of helium and quenched hydrogen , in the process of turning into normal da stars , and that the non - da gap could be accounted for by the sudden transformation of apparently helium - rich atmosphere white dwarfs into hydrogen - dominated atmospheres .
more recently , @xcite have explored more quantitatively the model proposed by brl .
the results of @xcite indicate that the accretion of even minute amounts of hydrogen onto a helium - dominated atmosphere , @xmath22 , provide enough free electrons through the thermal ionization of hydrogen to increase the he@xmath23 opacity significantly , which in turn , reduces the atmospheric pressure below which quenching effects on hydrogen become negligible .
the presence of hydrogen therefore has an _ indirect _ effect on the total opacity through the free electrons it provides to the plasma . on the basis of their analysis
, @xcite rejected the model proposed by brl .
given these results , the nature of the peculiar objects identified by brl and the existence of the non - da gap remain unexplained .
@xcite provided an alternative explanation for the non - da gap .
the gap is evident in his figure 13 even though he restricted his sample of objects to that of @xcite rather than using the more complete sample studied by brl , where the gap is even more obvious .
hansen has shown that the concentration of non - da stars just below @xmath15 k have nearly the same age as the concentration of da stars around 6000 k. on the basis of this conclusion alone , hansen suggested that the non - da gap alluded to by brl is only a result of the difference of cooling timescales between hydrogen - dominated and helium - dominated atmospheres . however
, this interpretation can not be correct since it neglects the fact that non - da stars are also observed in large numbers _ above _ the gap at @xmath24 k. in other words , non - da stars are seen both above and below the gap , and the high - temperature end of the gap does not reflect a simple termination of the cooling sequence of degenerates with helium - dominated atmospheres . unless helium - rich white dwarfs evolve more rapidly _ through _ the gap than above or below the gap , which is not the case , an explanation based solely on differences in cooling rates does not hold .
obviously , a chemical evolution from one spectral type to another must be invoked .
other results from the brl analysis support the idea that some physical mechanism , yet unidentified , is at play near the blue edge of the non - da gap .
for instance , the most widely accepted explanation for the presence of carbon in the photospheric layers of dq stars is that carbon is convectively dredged - up from the core , as demonstrated quantitatively by @xcite .
these calculations show that the extent of the carbon diffusion tail is reduced at low effective temperature due to the partial recombination of carbon , and smaller carbon abundances are predicted in dq stars with lower effective temperature , in apparent agreement with the observations ( see fig . 4 of * ?
as pointed out by brl , however , the coolest dq stars have the strongest observed c@xmath5 molecular bands , and cooler objects with even smaller abundances are well within the observational limits and should be easily detectable .
the brl analysis has revealed instead the existence of an abrupt cutoff in effective temperature at @xmath25 k below which dq stars are no longer found .
the presence of carbon in stars below this temperature threshold is in the form of what has been interpreted as c@xmath5h molecular bands by @xcite .
the most obvious explanation proposed by brl is that dq stars turn into c@xmath5h stars below @xmath26 k. the mechanism proposed by brl by which the hydrogen required to form the c@xmath5h molecule suddenly appears in these objects relies on their composition transition model , which was later discarded by @xcite .
hence , even though the dq to c@xmath5h transition remains the most likely explanation for the existence of the abrupt cutoff of the dq star phenomenon , the physical mechanism by which this transition occurs needs further investigation .
brl have also shown that below the non - da gap near @xmath15 k , non - da stars reappear in large number .
the most likely explanation for this phenomenon is the convective mixing of the relatively thin surface hydrogen layer with the deeper and more massive helium envelope .
detailed envelope calculations indicate that the bottom of the hydrogen convection layer reaches its deepest point at a certain temperature ( see , e.g. , * ? ? ?
* ) ; white dwarfs with hydrogen layers thicker than this maximum depth point can never mix and will remain hydrogen - rich forever .
the canonical value of the lowest possible mixing temperature is @xmath27 k , which corresponds to a hydrogen layer mass of @xmath28 .
this is @xmath29 k hotter than the temperature required to account for the brl observations .
new envelope models discussed by brl which include realistic model atmospheres as boundary conditions reduce the mixing temperature limit well down to 5000 k , however .
hence , convective mixing may well represent the correct interpretation for the reappearance of non - da stars below the non - da gap .
the continued presence of da stars requires that at least some stars have hydrogen layers thicker than @xmath28 . following the brl analysis ,
@xcite have gathered new photometric and spectroscopic data for the 43 cool , low - luminosity , high proper - motion white dwarfs in the @xcite sample .
this sample was analyzed with the techniques developed by brl in order to improve the determination of the cool white dwarf luminosity function and the estimate of the age of the local galactic disk .
brl discuss at length the shortcomings of the previous estimates by @xcite , and in particular the lack of precise atmospheric parameter determinations as well as accurate bolometric corrections for the stars in the sample . the improved luminosity function of @xcite yields an age of the local region of the galactic disk of @xmath30 gyr , with the uncertainty being mainly due to the core composition , the effect of separation on crystallization , and the unknown mass of the outer helium envelope .
the @xcite sample is being extended to a lower proper - motion limit and a new luminosity function is being determined which supports the earlier work @xcite . as mentioned in brl ,
these earlier efforts represent only the first step towards a better understanding of the chemical evolution of cool white dwarfs .
even though the brl analysis has provided us with a wealth of new observational and theoretical results , many questions have remained unanswered , and the problems to be solved have multiplied .
the success of our analysis has prompted us to extend our study to a larger and more complete sample of cool white dwarfs .
we thus present in this paper a detailed analysis of all cool ( @xmath0 k ) white dwarfs with trigonometric parallax measurements available to us .
such measurements are of particular importance to derive stellar radii , and hence stellar masses .
this new sample is comprised of 152 white dwarfs for which masses can be derived .
in contrast , only 67 out of the 110 white dwarfs analyzed by brl had trigonometric parallax measurements .
furthermore , brl concentrated their effort towards cool ( @xmath31 k ) , non - da white dwarfs , while our new sample is distributed over a wider range of effective temperature and spectral types . in 2 ,
we discuss the improvements in the model atmospheres and cooling models since our most recent analysis . in 3 , we present our new photometric and spectroscopic observations , which , combined with our earlier observations , are analyzed in 4 .
the global properties of our sample are examined in 5 .
our conclusions follow in 6 .
the model atmospheres used in this analysis are described at length in @xcite , and improvements to these models are discussed in brl and below .
these models are appropriate for cool white dwarfs with pure hydrogen or pure helium atmospheres , as well as mixed h / he compositions .
our current model grid now includes self - consistently the @xmath32 ion in the ionization equilibrium calculations . as discussed in detail by @xcite ,
the @xmath32 ion was treated as a trace element in our earlier calculations .
although this assumption is valid in terms of the number of particles , @xmath33 , the presence of the @xmath32 ion affects the ionization equilibrium significantly for temperatures below @xmath34 k. in cool , pure helium ( or helium - rich ) models , the electron density is governed by he@xmath35 , rather than he@xmath36 ( or @xmath21 ; see @xcite ) .
hence , despite the fact that the continuous absorption opacity of he@xmath35 is negligible at low temperatures , the inclusion of this ion in model atmospheres has repercussion on the he@xmath23 free - free opacity the dominant opacity source in helium - rich models through its effect on the electron density .
the occupation probability of @xcite for the calculation of the hydrogen populations has also been included more accurately in our current model grid following the work of @xcite .
this formalism allows us to calculate cool white dwarf models with arbitrarily low hydrogen abundances .
indeed , the photospheric pressures which characterize low hydrogen abundance atmospheres are so high that the hydrogen atomic levels ( including that of the @xmath21 ion ) are strongly perturbed . in the most extreme cases ,
pressure ionization of hydrogen may even occur ( we note that even though the pressure ionization model used in the hummer - mihalas formalism is very crude , it is sufficient for our current needs ) . since the hummer - mihalas formalism had not been included in our earlier investigations , brl were unable to compute cool models with @xmath37 ( see 5.3 of brl ) .
we note that the nonideal effects included in the equation - of - state of our pure helium model grid ( see * ? ? ? * and references therein ) are neglected in the mixed h / he models .
our latest hydrogen ( and mixed hydrogen / helium ) models include the new h@xmath5 collision - induced opacity calculations of @xcite , as well as the additional term discussed by @xcite which takes into account collisions involving three h@xmath5 molecules .
our hydrogen model grid has also been extended to @xmath38 k , the coolest models calculated by @xcite .
the non - ideal effects in the equation - of - state introduced by saumon and jacobson are neglected here , however .
a more detailed discussion of these cooler models will be presented elsewhere .
the hydrogen and mixed hydrogen / helium models have been calculated with the so - called ml2/@xmath39 parameterization of the mixing - length theory ( as opposed to the ml1 calibration used in brl ) , following the suggestion of @xcite from their analysis of zz ceti stars .
although this will not affect the atmospheric parameter determinations of the cooler objects in our sample ( @xmath40 k ) , the emergent energy distributions of warmer models are quite sensitive to the convective efficiency used in the calculations , as shown by @xcite .
we note finally that this new model grid has not changed significantly the quality of the fits shown in brl , or the determination of the atmospheric parameters from our previous analyses .
recently , @xcite have studied the importance of including metals in the atmospheres of cool , helium - rich white dwarfs .
their results indicate that the model atmospheres and predicted energy distributions are quite sensitive to the presence of additional elements .
they also claim that their improved models reproduce the scatter observed in two - color diagrams better than the models used in brl .
here , we present our own assessment of this problem . a more complete appraisal of the jrgensen et al .
results will be presented elsewhere @xcite .
the main effect of including heavy elements in the model atmospheres of cool , helium - rich white dwarfs is to increase the number of free electrons in the plasma , which in turn affect the he@xmath23 free - free opacity considerably . as mentioned by @xcite , more than 99% of the electrons may come from metals which have a relatively low ionization potential . as an experiment
, we calculated a pure helium model at @xmath41 k , @xmath42 , but with an electron density artificially increased by a factor of 1000 ( only 0.1% of the electrons come from helium directly , i.e. the value quoted by jrgensen et al . ) .
the temperature and pressure structures of this model are compared in figure [ fg : comp_tp ] with those of a standard pure helium model . as can be seen ,
the temperature structure of the upper layers has changed significantly , but that of the deeper layers where the _ continuum is formed _
( @xmath43 ) has remained relatively unaffected .
the gas pressure , on the other hand , has been reduced quite dramatically throughout the atmosphere as a result of the increase in the opacity ( @xmath44 ) , which in this case is dominated by the he@xmath23 free - free opacity .
perhaps of greater interest is the comparison of the predicted fluxes from these two models , displayed in figure [ fg : compflux ] .
not surprisingly , the results indicate that the observed differences are small .
the reason for this is that in the pure helium model , the he@xmath23 free - free opacity already dominates at all wavelengths where the continuum is formed .
an increase in the electron density thus increases the opacity by the same factor _ at all wavelengths_. consequently , the emergent flux can not be redistributed any differently .
the small differences observed in figure [ fg : compflux ] can be measured more quantitatively by fitting our test model with our pure helium grid .
we find that the predicted energy distribution of our test model can be fit to better than 1% at all wavelengths with our pure helium grid at @xmath45 k. we thus conclude that the presence of metals in helium - rich models does not affect significantly the atmospheric parameters derived in the brl analysis and in this analysis .
note that the presence of metallic lines or molecular absorption features may affect the predicted broadband fluxes , but we do not believe that they can affect the model structure significantly , with the exception of the most extreme cases such as those discussed in [ sbsc : optir ] .
there are several reasons why the conclusions of @xcite are different from those reached here .
first they state that white dwarfs are `` carbon rich '' .
this is simply not true as only a small fraction of cool non - da stars exhibit carbon features .
it is also a well known observational fact that the presence of carbon and metallic features are mutually exclusive ( for reasons still unknown ) .
the authors have also shown in their figure 3 that the inclusion of molecules in their model produces a substantial heating at all layers . however ,
this statement is somewhat misleading since their plot shows the temperature profile as a function of gas pressure , and as these two parameters are not independent in a model atmosphere , it is not possible from such a plot to determine which of the parameters ( temperature or gas pressure ) has in fact changed .
a more appropriate independent parameter is the rosseland mean optical depth , as used here in figure [ fg : comp_tp ] .
jrgensen et al .
also claim that the new h@xmath5-he collision - induced opacity they have calculated and not used in our own model calculations has a significant influence on the predicted emergent fluxes .
however , their figure 6 shows instead that the differences are actually quite small at the @xmath46 values considered here , and not pertinent to the present discussion .
more importantly , they have assumed a non - zero hydrogen abundance of @xmath47 on the basis of either old studies , or analyses of objects that do show direct or indirect traces of hydrogen ( e.g. , ross 640 , lhs 1126 ) .
traces of hydrogen in these cool models produce a strong infrared flux deficiency ( see their fig .
6 ) which is simply not observed in cool non - da stars studied by us ( brl and this analysis ) , with the glaring exception of lhs 1126 @xcite near @xmath48 k whose infrared flux deficiency has been attributed to the h@xmath5-he molecular opacity .
surprisingly enough , @xcite conclude that at this particular temperature , the h@xmath5-he opacity has little effect .
this is because they have considered hydrogen abundances which are way too small in this case lhs 1126 has @xmath49 .
the claim by jrgensen et al . that the predicted energy distribution is not sensitive to the helium - to - hydrogen abundance ratio is thus also inaccurate .
their figures 8 and 9 show their helium - rich models , as well as some of their hydrogen - rich models , in a ( @xmath50@xmath51 , @xmath51@xmath52 ) two - color diagram , together with the brl observations .
also shown for comparison are the hydrogen - rich and helium - rich models used in brl .
the authors conclude that both their hydrogen - rich and helium - rich models reproduce the observed photometric data better than the models used in brl . however , the authors failed to distinguish the da and non - da stars in their figures , as was carefully done in figure 9 of brl .
a more careful examination of both results indicates that their helium - rich models actually fail to match the non - da stars close to @xmath50@xmath53 , @xmath51@xmath54 altogether , while our pure helium models go straight through the observed data .
their figure 8 shows instead that their helium - rich models reproduce nicely the sequence of _ da stars _ which have hydrogen - rich atmospheres .
the authors also claim that at the warm end of the observed sequence , their models `` follow better the tendency of a linear observed sequence that the observations indicate , than do the brl models which diverge toward higher values of @xmath50@xmath51 '' .
again , the authors failed to realize that the warm end of the sequence in brl is composed exclusively of non - da stars .
the fact is that da stars _ do diverge _ toward higher values of @xmath50@xmath51 , as do the models ( see our figure [ fg : bmv : vmk ] below ) .
if anything , their hydrogen - rich models fail to reproduce the warm end of the observed da sequence .
finally and more importantly , @xcite claim that their helium - rich models reproduce the scatter in the observational data seen at the cool end of diagram better than do the brl models which follow an almost straight line .
first of all , most of the observed scatter is due to the fact that the coolest objects include _ both hydrogen - rich and helium - rich white dwarfs_. indeed , when one considers only the helium - rich stars ( the filled circles in the right panel of fig .
9 of brl ) , they do form an almost straight line , as opposed to the jrgensen et al .
second , these authors show that their models can reproduce the entire cool end of the diagram by changing the c / o ratio .
this is again somewhat misleading .
indeed , the observed `` hook '' in their predicted colors is entirely due to the h@xmath5-he collision - induced opacity .
similar hooks can be observed in the brl models as well , whether they are due to h@xmath5-h@xmath5 or h@xmath5-he molecular opacities ( see fig .
9 of brl ) .
since these collision - induced opacities are strongly dependent on the atmospheric pressure , the predicted colors will be sensitive to surface gravity ( see the hydrogen sequences with various values of @xmath55 in fig .
9 of brl , middle panel ) , or to the hydrogen - to - helium abundance ( see the helium - rich sequences with various values of @xmath56 in fig . 9 of brl , right panel ) .
the results shown in figure 8 of @xcite simply reflect a change in the opacity , and thus atmospheric pressure , due to a change in the electron density for various values of c / o .
more problematic is the effective temperature near 3000 k inferred from figures 8 and 9 of @xcite for the bulk of the coolest white dwarfs . at this temperature ,
the predicted energy distribution shows a strong infrared flux deficiency which is simply not observed in these objects , as discussed above .
this is the danger of interpreting cool white dwarfs in two - color diagrams only , and in particular the ( @xmath50@xmath51 , @xmath51@xmath52 ) diagram ( see 5.2.2 of brl ) , rather than fitting the entire energy distribution , which is the technique favored by brl and in this analysis . in order to derive stellar masses from the measured radii , and to determine cooling ages
, we must rely on detailed evolutionary models .
in brl , we made use of the models of @xcite for the pure helium and mixed hydrogen / helium compositions , and those of @xcite for the pure hydrogen models .
the former models have carbon - core compositions , helium layers of @xmath57 , and no hydrogen layers , while the latter have carbon - core compositions as well , but helium layers of @xmath58 , and thick hydrogen layers of @xmath59 . in this paper
, we use instead two families of new evolutionary models kindly computed for us by g. fontaine who used the new evolutionary code developed by p. brassard at universit de montral on the basis of finite - element techniques . while details of that code will be described elsewhere @xcite , a few remarks about it are worthy of interest .
the code is especially designed to follow in a robust and efficient way moving discontinuities in a cooling white dwarf model , namely , the top and base of superficial convection zones , and the advancing crystallization front in the core and in the envelope .
the full structure of a model , from the center to the top of the atmosphere ( typically located at @xmath60 = @xmath61 ) , is included in the evolutionary calculations . upgraded input physics , including the new model atmospheres described above , has been incorporated in the code in order to be able to compute realistic white dwarf models reaching the regime of very low effective temperatures that is of interest for discussing white dwarf populations in the halo and in old globular clusters .
the first family of evolutionary structures that we use are standard `` thick envelope '' stellar models with a helium mantle of @xmath62 , and an outermost hydrogen layer of @xmath59 .
the core of these models consists of a uniform mixture of carbon and oxygen in equal proportions ( @xmath63 ) .
no sedimentation between carbon and oxygen upon crystallization has been taken into account in these models .
sedimentation would add a small delay to the cooling age , but would not change the mass - radius relation in any significant way . on the other hand , the effects of convective feedback on the position of the hydrogen convection zone have been taken into account in the calculations .
likewise , neutrino cooling has been included in a routine way , but this only affects the evolution in the high luminosity phases only , as is well known .
no residual thermonuclear burning has been included in order to keep the same value of the fractional mass of the outer hydrogen layer in all models .
finally , 16 different masses in the range 0.21.3 @xmath6 have been considered .
the second family of models ( with again 16 different sequences ) is similar to the first one , except for the assumed stratification of the envelope .
this second batch of structures are now `` thin envelope '' stellar models with a helium mantle of @xmath58 as before , but with an outermost hydrogen layer of only @xmath64 .
no convective mixing between the hydrogen and helium layers is allowed in these models .
we picked this particular stratification as representative of `` thin '' hydrogen envelope models .
the small layer of hydrogen atop these models does not change in any significant way the mass - radius relation for the cooler models that are of interest here . from that point of view
, that same mass - radius relation can thus be used for non - da stars .
evolutionary calculations of helium - rich atmospheres can also be carried out , but the cooling times show an extreme sensitivity to the presence of even very small traces of heavy elements in the atmosphere / envelope , leading to a continuum of ages , hence we adopted the current strategy .
in the brl study , 110 genuine cool white dwarfs were selected on the basis of their spectral type classification from the catalog of @xcite , from a proper - motion survey in the southern hemisphere @xcite , and from the study of @xcite .
this first sample was strongly biased against da stars in order to determine whether true dc stars existed , and if they did , what was their atmospheric composition .
our high signal - to - noise spectroscopy actually revealed the presence of h@xmath3 in 20 of the previously classified non - da stars , a result which stresses the importance of large aperture telescopes for studying cool white dwarfs , as first emphasized by @xcite .
several additional da , dq , and dz stars were included in the sample as well to test the validity of our model atmosphere calculations , and to improve our understanding of the chemical evolution of cool white dwarfs .
brl found that even though the model energy distributions are somewhat sensitive to the surface gravity , it is practically impossible to determine @xmath55 from the observed photometry alone . only for stars with available trigonometric parallax measurements
is it possible to determine the stellar radius , and thus the mass through the mass - radius relation . in the brl sample ,
67 white dwarfs had trigonometric parallax measurements taken from the yale parallax catalog ( * ? ? ?
* hereafter ypc ) , and from @xcite , @xcite , @xcite , and @xcite .
stellar masses were determined for a subset of 60 white dwarfs ( 6 objects had unreliable parallaxes , while one dq star had carbon bands too strong to be analyzed properly ) .
brl assumed a surface gravity of @xmath65 for the remaining objects in their sample .
the natural step following the comprehensive study of brl was to gather spectroscopic and photometric data for all cool white dwarfs with measured trigonometric parallaxes available in the literature .
a total of 161 da and non - da white dwarfs with a temperature index ( @xmath66 ) of 4 and cooler were thus selected from the ypc , and from @xcite .
this temperature cutoff corresponds roughly to that below which convective mixing of da stars is believed to occur .
only stars with parallax uncertainties smaller than 30% were considered ; 7 objects were rejected on this basis ( 0145@xmath67174 , 1042@xmath68593 , 1639@xmath68153 , 1708@xmath67147 , 2139@xmath68132 , 2240@xmath67017 , and 2254@xmath68076 ) .
two other objects were not considered since one is a known dc + m dwarf system ( 1133@xmath68358 ) , while the other one has a known spectroscopic temperature in excess of 20,000 k ( 1919@xmath68145 ) .
hence , 152 white dwarf candidates were retained for photometric and spectroscopic observations . among this sample ,
67 objects are in common with brl , and new photometric and spectroscopic data are presented in this paper for 85 white dwarfs .
since this new sample is based on trigonometric parallax measurements , it should contain no bias against any specific spectral type , in contrast with the brl analysis which favored non - da white dwarfs .
the number of da and non - da stars as a function of @xmath46 for our trigonometric parallax sample is displayed in figure [ fg : davsnonda ] .
also shown is the subset of objects in the brl sample whose stellar masses could be determined from trigonometric parallax measurements @xmath69 .
the temperature estimates are obtained from the detailed fits to the photometric energy distributions ( 5 of brl and [ sbsc : adopt ] below ; 60 stars in the brl sample and 150 stars in this analysis ) , while the spectral types are determined from our high signal - to - noise spectroscopic observations .
the comparison indicates that our new sample has significantly increased the number of da stars above @xmath70 k and the number of non - da stars at all temperatures with respect to the brl sample . also apparent from this figure
is the paucity of non - da stars in the 5000 - 6000 k temperature range .
note that the coolest bin of the non - da panel may contain a significant number of hydrogen - rich white dwarfs that are simply too cool to exhibit an h@xmath3 feature .
high signal - to - noise spectroscopy has been obtained for all 152 objects in our sample , 15 of which have been secured by @xcite .
some of the spectroscopic observations have already been presented in brl and in @xcite .
the new spectra reported here have been obtained during several observing runs using the kpno and ctio 4 m telescopes equipped with the rc spectrograph ( 1996 , 1998 ; where the years indicate the period the observations were obtained ) , and the eso 3.6 m telescope equipped with the efosc ( 1996 , 1998 ) .
the spectral resolution varies between 5 and 15 ( 18 for the greenstein spectra ) . details of our observing procedure and data reduction are provided in @xcite .
the spectral coverage varies according to the detector used although all spectra cover at least the h@xmath3 region . the new spectra not already shown in brl which show h@xmath3 are displayed in figure [ fg : plotspecda ] in order of decreasing equivalent widths .
the names of the 7 newly identified da stars are italicized in figure [ fg : plotspecda ] .
we note that the coolest da stars in our sample are all new da identifications , a result which stresses once again the importance of large aperture telescopes for the spectral classification of cool white dwarfs ( brl , * ? ? ?
objects of particular interest in figure [ fg : plotspecda ] include two magnetic white dwarfs with discernible zeeman splitting : gd 175 ( 1503@xmath67070 ) discovered in this survey , and g128@xmath6772 ( 2329 + 267 ) first identified by @xcite .
there are also two objects with particularly broad and shallow profiles : the dza stars ross 640 ( 1626@xmath68368 ) and the dzqa star l745@xmath6746a ( 0738@xmath67172 ) .
both of these objects have helium - dominated atmospheres ( * ? ? ?
* this analysis ) , with their h@xmath3 line profile broadened by van der waals interactions with neutral helium ; we note that such objects are rare in our sample .
the h@xmath3 line profile of the g107@xmath6770a / b system ( 0727@xmath68482a / b ) is also broad and shallow with respect to other cool objects . in this case , however , since the a and b components of the system are not resolved , it is more likely that we are dealing with a normal da star whose line profile is diluted by a featureless white dwarf companion ( see [ sbsc : addobj ] , however ) .
the central component of the zeeman triplet in emission is also clearly visible in the spectrum of gd 356 ( 1639@xmath68537 ; see also * ? ? ?
* ) , while the @xmath71 components are outside the wavelength range displayed here .
finally , the coolest da star shown in figure [ fg : plotspecda ] is g227@xmath6728 ( 1820@xmath68609 ) with an effective temperature of 4780 k. this is still 500 k hotter than the coolest da star in our complete sample , lhs 239 ( 0747 + 073b ) , with @xmath72 k ( see brl and [ sbsc : adopt ] below ) . our non - da spectra
not already shown in brl are displayed in figure [ fg : plotspecnonda ] .
as noted in brl , a featureless white dwarf spectrum does not necessarily imply a helium - rich atmosphere since a hydrogen - rich white dwarf may be too cool to show an h@xmath3absorption feature , and a full analysis of the energy distribution is required to establish properly the atmospheric composition of these objects .
the subset of dq stars with blue spectral coverage is displayed in figure [ fg : plotspecdq ] in order of decreasing effective temperature ( as determined in
[ sbsc : adopt ] ) .
these are all well known dq stars with the exception of two new identifications , lhs 2392 ( 1115@xmath67029 ) and g184@xmath6712 ( 1831@xmath68197 ) . as discussed in brl and above
, the convective dredge - up model of @xcite predicts a continuous decrease of carbon abundances with decreasing effective temperature .
as seen from figure [ fg : plotspecdq ] , however , the coolest dq stars have the strongest carbon features , and therefore , the sudden disappearance of the dq phenomenon at even cooler temperatures needs to be attributed to an additional effect .
optical @xmath1 and infrared @xmath2 photometry has been secured for 150 and 148 program objects , respectively , out of our complete sample of 152 white dwarfs .
since some optical photometry is required for fitting the energy distribution , @xmath73 photometry for the 2 missing objects 0326@xmath67273 and 0839@xmath67327 has been taken from @xcite .
most of the data for the 67 objects in common with brl were acquired between 1991 august and 1996 february , but we have supplemented that data with some newer photometry .
additional @xmath1 photometry presented here has been obtained with the 0.9 m telescopes at kpno ( 1996 , 1998 ) and ctio ( 1995 ) , while the new @xmath2 photometry has been obtained with the ctio blanco 4 m telescope using cirim ( 1995 ) , the nsfcam camera on the nasa irtf ( 1996 ) , the ircam on ukirt ( 1996 , 1998 ) , and ukirt service time ( 1997 to 1999 ) . details of our observing procedure and data reduction are provided in @xcite .
the @xmath1 and @xmath2 photometry is presented in table 1 together with the number of independent observations ( @xmath74 ) .
also given in table 1 are the wd number and the name of each object , the measured trigonometric parallax and uncertainty , and the h@xmath3 equivalent width ( @xmath75 ) .
a value of @xmath76 implies a featureless spectrum near the h@xmath3 region ; for the magnetic da stars , the equivalent width is taken as the sum of the three zeeman components .
the photometric uncertainties are typically 3% at @xmath51 , @xmath77 , and @xmath78 , and 5% elsewhere , with the exception of the data marked `` : '' which indicates a 10% uncertainty .
the optical and infrared photometry is on the cousins system @xcite and the cit system @xcite , respectively .
the optical and infrared photometry of the a and b components of the g107@xmath6770 ( 0727@xmath68482 ) system is based on our combined optical photometry and the @xmath79 mag difference reported by @xcite and references therein , as well as on our own magnitude difference measurement of @xmath80 mag at @xmath81 , @xmath82 , and @xmath52 .
note that the spectrum of g107@xmath6770a / b displayed in figure [ fg : plotspecda ] is the _ combined _ spectrum of both components .
the @xmath12 vs ( @xmath51@xmath78 ) color - magnitude diagram is displayed in figure [ fg : mv : vmi ] for 149 stars ; 3 objects in table 1 have no ( @xmath51@xmath78 ) measurement .
the left panel shows all objects together with the error measurements associated with the trigonometric parallax measurements . in the middle and right panels ,
the same objects are distinguished in terms of their da or non - da spectral types .
the pure hydrogen and pure helium cooling sequences , calculated as in @xcite but with our latest model grid and cooling models discussed in [ sbsc : thfr ] , are superimposed on the observed data .
da and non - da stars form well - defined narrow sequences in this diagram , although not as narrow as the usno data set of @xcite , also shown in @xcite , which is a much more homogeneous parallax sample than the one studied here .
interestingly , all overluminous stars with apparent masses below 0.4 @xmath6 are of the da spectral type , with the exception of lp 31@xmath6740 ( 0324@xmath68738 ) which is probably too cool to show h@xmath3 anyway . as discussed in brl , most , if not all , of these objects are unresolved binaries ( e.g. l870@xmath672 , * ? ? ?
* ) and their luminosity is the contribution of two white dwarfs with probably normal masses .
the color - magnitude diagram presented in figure [ fg : mv : vmi ] indicates that such systems are mostly composed of da white dwarfs , a result which suggests a particular evolutionary path . we will come back to this point below ( see [ sbsc : mdist ] ) .
a particularly interesting object in figure [ fg : mv : vmi ] is the non - da star eso 439@xmath6726 ( 1136@xmath67286 ) analyzed in detail by @xcite .
this object with its absolute visual magnitude of @xmath83 was first interpreted as an extremely old and thus cool white dwarf .
however , the @xcite analysis convincingly demonstrated that the low luminosity of this object could be attributed to its small radius ( or high mass ) .
the most recent parameters derived for this object come from the brl analysis , @xmath84 k and @xmath85 @xmath6 .
however , as first demonstrated by @xcite , white dwarfs with @xmath86 k become increasingly _ bluer_. an inspection of figure [ fg : mv : vmi ] now suggests an alternate interpretation for eso 439@xmath6726 , namely that this star could be a @xmath87 k , @xmath88 @xmath6 white dwarf ; such a low temperature would imply an age of roughly 13 gyr , i.e. twice as old as the age determined by brl for this object .
the detailed analysis of the entire energy distribution presented below will show that the infrared @xmath81 magnitude is totally inconsistent with this last interpretation , and that the earlier conclusions reached by @xcite are still valid .
if the overluminous da stars and eso 439@xmath6726 are excluded , all objects in figure [ fg : mv : vmi ] have masses in the range 0.4 0.5ex@xmath89@xmath6 0.5ex1.0 .
if we assume that most da stars have hydrogen - rich atmospheres , as demonstrated by brl , their mean mass inferred from figure [ fg : mv : vmi ] is somewhere around 0.7 @xmath6 .
similarly , if we assume that most non - da stars have helium - rich atmospheres ( with the exception of the hydrogen - rich white dwarfs which are too cool to show h@xmath3 ) , their mean mass inferred from figure [ fg : mv : vmi ] is also somewhere near 0.7 @xmath6 .
both of these average masses are about 0.1 @xmath6 higher than those inferred from the spectroscopic analyses of warmer da and db stars @xcite . under similar assumptions ,
the coolest da and non - da stars in figure [ fg : mv : vmi ] have temperatures near @xmath90 k based on their location in this diagram with respect to the pure hydrogen and pure helium cooling sequences , respectively . as mentioned above , however
, the coolest non - da stars may well have a hydrogen - rich atmospheric composition , and only a complete analysis of the combined energy distribution will yield reliable temperature estimates . as can be seen from figure [ fg : mv : vmi ] , such an uncertainty in the atmospheric composition could easily result in a @xmath91 k temperature uncertainty for the coolest non - da white dwarfs in our sample .
the ( @xmath50@xmath51 , @xmath51@xmath52 ) two - color diagram is displayed in figure [ fg : bmv : vmk ] for 145 objects ; 2 objects in table 1 have no ( @xmath50@xmath51 ) measurement while 5 have no @xmath52 measurement . da and non - da stars form two distinct sequences which overlap near ( @xmath51@xmath52)@xmath92 . this behavior
is observed in the model sequences as well .
we note in particular that both the observed and predicted da sequences tend to diverge toward higher values of @xmath50@xmath51 for @xmath51@xmath93 , in contrast with the models of @xcite which follow instead the non - da sequence in this range of @xmath51@xmath52 .
even though the observed da sequence is well reproduced by the 0.6 @xmath6 pure hydrogen models at the hot end of the sample , an increasing departure is observed above ( @xmath51@xmath52)@xmath94 . in these regions the cool observed da sequence is better reproduced by the pure helium models !
brl proposed an explanation for this departure in terms of a missing opacity source near the @xmath50 filter in the pure hydrogen models , most likely due to a pseudocontinuum opacity originating from the lyman edge ( see 5.2.2 of brl for a complete description ) .
figure 13 of brl shows indeed how the agreement between the observations and the model predictions could be improved by including , even approximately , such an opacity in the model calculations .
however , the lack of a more precise theoretical framework precludes us from including this opacity in our current model grid , and until this can be accomplished , we must refrain from interpreting such ( @xmath50@xmath51 , @xmath51@xmath52 ) diagrams for da stars any further . on the other hand , the non - da observed sequence in figure [ fg : bmv : vmk ] and the pure helium models are in much better agreement .
non - da stars , most of which have helium - rich atmospheres , are characterized by more transparent atmospheres than hydrogen - rich atmospheres , and as a result , their broadband colors are particularly sensitive to the presence of strong metallic lines or molecular bands . consequently , several non - da stars in figure [ fg : bmv : vmk ] lie well outside the `` mean '' observed sequence .
this is the case for lp 701@xmath6729 ( 2251@xmath67070 ) whose spectrum exhibits a strong drop in the flux shortward of @xmath95 attributed to ca i and ii lines @xcite , lhs 1126 ( 0038@xmath67226 ) which shows a strong infrared flux deficiency attributed to h@xmath5-he collision - induced absorptions @xcite , g165@xmath677 ( 1328@xmath68307 ) a heavily blanketed white dwarf with strong metallic lines @xcite , g240@xmath6772 ( 1748@xmath68708 ) whose spectrum is characterized by a `` yellow sag '' @xmath96% deep and @xmath97 wide ( see , e.g. , fig .
19 of * ? ? ?
* ) , and the strong dq star bpm 27606 ( 2154@xmath67512 ) whose spectrum is displayed in figure [ fg : plotspecdq ] .
because of the theoretical problems outlined above with the pure hydrogen models , brl preferred to rely on the ( @xmath51@xmath78 , @xmath51@xmath52 ) two - color diagram in which a much better agreement between theory and observation could be achieved
. such a diagram is displayed in figure [ fg : vmi : vmk ] for 144 objects ; 3 objects in table 1 have no ( @xmath50@xmath51 ) measurement while 5 have no @xmath52 measurement .
the results show that both da and non - da stars form narrow cooling sequences , but that they are offset by @xmath98 mag .
the outliers are all of the non - da type , some of which have been identified in the right panel and already discussed above .
g148-b4b ( 1215@xmath68323 ) has most likely its @xmath78 magnitude contaminated by the close m dwarf companion , while g195@xmath6719 ( 0912@xmath68536 ) is a @xmath99 mg magnetic white dwarf with some undetermined spectroscopic features @xcite .
the non - da stars lp 131@xmath6766 ( 1247@xmath68550 ) and er 8 ( 1310@xmath67472 ) actually have hydrogen - rich atmospheres ( see brl and the detailed analysis below ) .
also shown in the da and non - da panels are the model sequences for pure hydrogen , pure helium , and mixed he / h atmospheres .
the departure of the pure hydrogen and mixed he / h models from the pure helium models is due to the collision - induced absorption by molecular hydrogen due to collisions with h@xmath5 and neutral helium , respectively . because of these strong infrared absorption features , the ( @xmath51@xmath78 , @xmath51@xmath52 ) two - color diagram is particularly useful for detecting extremely cool hydrogen - rich white dwarfs , as well as cool white dwarfs with mixed hydrogen and helium atmospheres .
as can be seen from figure [ fg : vmi : vmk ] , such objects are extremely rare , with the glaring exception of lhs 1126 which has already been analyzed in brl ( @xmath100 k , @xmath101 ; see their fig .
31 ) . finally , the existence of a non - da gap in the 5000 - 6000 k temperature range is already obvious in this diagram , especially when the non - da sequence is contrasted with the da sequence .
the only two objects inside the gap are lhs 2710 ( 1313@xmath67198 ) and eso 292@xmath6743 ( 2345@xmath67447 ) which have already been analyzed by brl and identified as being peculiar ( see 6.3.1 of brl for more details ) .
it is again unlikely that the explanation proposed by @xcite , namely that different evolutionary timescales of hydrogen - rich and helium - rich atmosphere white dwarfs , can account for the _ observational _ fact shown in the right panel of figure [ fg : vmi : vmk ] , unless non - da stars evolve more rapidly within the gap than outside the gap .
our fitting technique is described at length in brl .
briefly , magnitudes are converted into observed fluxes , and the resulting energy distributions are fitted with those predicted from our model atmosphere calculations using a nonlinear least - squares method . only @xmath46 and the solid angle @xmath102 are considered free parameters .
the distance @xmath103 is obtained from the trigonometric parallax measurement , and the stellar radius @xmath77 is converted into mass using the new cooling sequences described in
[ sbsc : thfr ] with thin hydrogen layers for the pure helium and mixed hydrogen / helium compositions , and those with thick hydrogen layers for the pure hydrogen models ; masses derived from thin hydrogen layer are about 0.03 @xmath6 smaller than those obtained from thick hydrogen models @xcite . as discussed in 3.3
, there is a significant uv flux deficiency in hydrogen - rich stars , most likely due to a pseudocontinuum opacity originating from the lyman edge . as such
, the @xmath50 magnitude has been omitted in the fits of hydrogen - rich stars cooler than @xmath48 k. for identical reasons , it has also been omitted in the fits of strong dz stars such as vma 2 ( 0046@xmath68051 ) , g165@xmath677 ( 1328@xmath68307 ) , lp 701@xmath6729 ( 2251@xmath67070 ) , etc .
some objects in our sample possess strong molecular absorption features which are not taken into account in our model flux calculations .
such objects include the dq stars , the c@xmath5h stars lhs 1126 ( 0038@xmath67226 ) , lhs 290 ( 1043@xmath67188 ) , and g225@xmath6768 ( 1633@xmath68572 ) , as well as g240@xmath6772 ( 1748@xmath68708 ) whose spectrum is characterized by the yellow sag discussed above .
brl have shown that the fits to the energy distribution of dq stars are not affected significantly by the presence of the c@xmath5 swan bands , with the glaring exception of bpm 27606 ( 2154@xmath67512 ) , the strongest dq star known .
for some of the strongest c@xmath5h stars and for g240@xmath6772 , the molecular absorption features are so important that the quality of the fits is poor . in these cases , we rely on a technique used in brl where the optical spectrum of the star is normalized to a continuum set to unity , and folded with the grid of model spectra . the presence of the strong molecular bands is thus taken into account , at least in first order , in the calculations of the emergent flux distributions .
this approximate procedure improves the quality of our fits markedly .
several fits to cool white dwarfs with various effective temperatures , surface gravities , and chemical compositions have already been displayed in brl .
in particular , their 5.3 and 5.4 show how the combined photometric energy distributions together with the high signal - to - noise spectra at h@xmath3 could be used to determine , or at least constrain , the chemical composition of cool degenerates .
the most important discriminant between the hydrogen - rich and helium - rich energy distributions lies near the infrared @xmath82 bandpass where hydrogen - dominated atmospheres exhibit a local maximum in their emergent flux due to the bound - free h@xmath23 opacity threshold at 1.6 .
this bump in the emergent flux distribution is of course not observed in pure helium atmospheres .
the chemical composition effects on our fits will not be repeated here , and we simply refer the reader to figures 15 and 16 of brl where several illustrative examples are displayed . here
we present instead typical fits to hydrogen - rich stars , helium - rich stars , stars with mixed compositions , confirmed or suspected double degenerates , as well as other objects of astrophysical interest .
we attempt as much as possible to present fits for objects which have not already been analyzed in brl . in the following plots , the observed fluxes and corresponding uncertainties
are represented by error bars while the model fluxes are shown as filled circles .
the atmospheric parameters of each fit are indicated in each panel .
sample fits for da stars covering the effective temperature range of interest are displayed in figure [ fg : sampleda ] ; sample fits for hydrogen - rich non - da white dwarfs too cool to exhibit h@xmath3 have already been shown in figure 26 of brl .
the values of @xmath46 and @xmath55 given in each panel are obtained from fits with pure hydrogen atmospheric compositions .
the effects of using models with small traces of helium are discussed in detail in 5.3 of brl .
the spectroscopic observations at h@xmath3 are not used directly in the fitting procedure , but they serve as an internal check of our photometric solutions .
the theoretical line profiles are simply interpolated at the values of @xmath46 and @xmath55 obtained from the energy distribution fits and compared with the observed line profiles .
as can be seen from figure [ fg : sampleda ] , the observed energy distributions are well reproduced by our theoretical models throughout the temperature range of interest , and the predicted line profiles are in excellent agreement with the solution derived from the photometric observations .
ltt 4816 ( 1236@xmath67495 ; top object in fig .
[ fg : sampleda ] ) is a known zz ceti star for which @xcite obtained a spectroscopic solution of @xmath104 k and @xmath105 , in good agreement with the parameters derived here @xmath106 k and @xmath107 .
it is worth emphasizing that the atmospheric parameter determination of both analyses is based on completely different techniques .
it the former case , @xcite used fits to the balmer line profiles ( h@xmath108 to h8 ) , while here we rely on @xmath46 and @xmath55 estimates based on photometric observations and trigonometric parallax measurements . note that both analyses use the same calibration of the mixing - length theory in the model calculations , namely the so - called ml2/@xmath39 parameterization .
interestingly enough , our photometric fit to ltt 4816 based on ml2/@xmath109 models yields @xmath110 k and @xmath111 ( with the same uncertainties as above ) , while the spectroscopic fit to the balmer lines based on the same models yields @xmath112 k and @xmath113 ( see table 2 of * ? ? ?
hence , even though the photometric solution is not very sensitive to the value of @xmath3 ( i.e. the ratio of the mixing length to the pressure scale height ) , the spectroscopic solution differs by 940 k in @xmath46 and 0.15 dex in @xmath55 .
moreover , the internal consistency between the spectroscopic and photometric solutions is much better with ml2/@xmath39 than with ml2/@xmath109 models .
a similar conclusion was reached by @xcite based on a comparison of the atmospheric parameters of 22 zz ceti stars obtained from fits to the ultraviolet energy distributions and fits to the balmer line profiles .
our result thus brings additional support to the conclusion that the ml2/@xmath39 calibration of the mixing - length theory provides an excellent internal consistency between atmospheric parameters determined with various techniques over a wide range of wavelengths , which now extends from the ultraviolet to the infrared .
a similar example is g74@xmath677 ( 0208@xmath68396 ; middle panel of fig .
[ fg : sampleda ] ) , a well - studied daz white dwarf which has been re - analyzed recently by @xcite who presented an analysis of optical and ultraviolet spectroscopic observations .
their spectroscopic fit to the balmer lines ( see their fig .
4 ) yields @xmath114 k and @xmath115 , in excellent agreement with our photometric solution @xmath116 k and @xmath117 .
more comparisons between photometric and spectroscopic solutions are provided in
[ sbsc : compspec ] .
the coolest object in figure [ fg : sampleda ] is lhs 1801 ( 0551@xmath68468 ) with @xmath118 k. the fit to its energy distribution suffers from the pathological problem discussed above for the coolest hydrogen - rich stars in our sample in which some unknown pseudocontinuum opacity unaccounted for in our model calculations affect the emergent flux near the @xmath50 bandpass .
note that this problem does not occur for lhs 2522 ( 1208@xmath68576 ; also shown in fig .
[ fg : sampleda ] ) , a da star which is only 500 k hotter than lhs 1801 .
hence , the phenomenon begins to manifest itself in the intermediate temperature regime near @xmath48 k. as discussed in
[ sbsc : fittec ] , the @xmath50 filter has been dropped in all fits to hydrogen - rich stars below this temperature threshold .
as shown in figure [ fg : sampleda ] , the internal consistency achieved in this manner of the fit to the energy distribution and the h@xmath3 line profile is more than satisfactory .
sample fits for helium - rich , non - da stars covering the effective temperature range of interest are displayed in figure [ fg : samplenonda ] .
the values of @xmath46 and @xmath55 given in each panel are obtained from fits with pure helium atmospheric compositions .
the spectroscopic fits are of course not shown here since all objects are featureless near the h@xmath3 region ( see figure [ fg : plotspecnonda ] ) .
pure helium models provide excellent fits to the energy distributions of all of these objects in the @xmath46 range considered in this analysis , with the exception of g24@xmath679 ( 2011@xmath68065 ) which is discussed below . in particular , at low effective temperatures ( see , e.g. , lhs 542 ) , the problem observed in hydrogen - rich stars at the blue end of the energy distribution
is not observed in helium - rich stars .
this supports our conclusion that some unaccounted uv opacity from hydrogen is at the origin of this problem . the first object in figure [ fg : samplenonda ] is lp 257@xmath6728 ( 0802@xmath68386 ) , a hot dz star with a blue spectrum showing strong h and k lines @xcite .
lhs 43 ( 1142@xmath67645 ) is a weak dq star ( see figure [ fg : plotspecdq ] ) .
brl have discussed at length the validity of the pure helium models to analyze dz and dq stars despite the presence of strong metallic lines or carbon bands .
lhs 2333 ( 1055@xmath67072 ) is classified as a da white dwarf in @xcite , although a note added in their catalog reports the star to be featureless according to @xcite . our own spectroscopic observation which covers the range
@xmath119 confirms the results of schmidt & smith that lhs 2333 has a featureless spectrum a true dc star .
brl have shown that dc stars in this temperature range ( @xmath120 k ) usually belong to a group of peculiar objects whose energy distributions are better reproduced with pure hydrogen models rather than with pure helium models ( for reasons still unknown ) .
a good example of such a peculiar white dwarf is the next object shown in figure [ fg : samplenonda ] , g24@xmath679 ( 2011@xmath68065 ) , where the fits with both pure helium and pure hydrogen models are displayed .
lhs 2333 seems to be an exception to the rule , unless there is something even more peculiar with this particular object .
it is tempting to suggest that lhs 2333 may have recently changed spectral type from da to dc .
however , despite the three da spectral classifications of this object in the catalog of @xcite , only that of @xcite is an actual spectroscopic observation ( the other two references in the catalog do not report any new observation ) . given that this classification relies on old photographic plates , it is difficult to claim that the spectral type of this object has actually changed over the 30 year or so period . finally , the last object in figure [ fg : samplenonda ] , lhs 542 ( 2316@xmath67064 ) , is a cool dc white dwarf according to our own spectrum which reveals no significant absorption feature in the range @xmath43780 - 6890 .
there is little evidence for a large population of cool white dwarfs with mixed hydrogen and helium atmospheres , as first demonstrated by brl .
such stars would show up easily in ( @xmath50@xmath51 , @xmath51@xmath52 ) or ( @xmath51@xmath78 , @xmath51@xmath52 ) two - color diagrams due to the h@xmath5he collision induced opacity which produces a strong infrared flux deficiency .
the c@xmath5h stars also have energy distributions characterized by such an infrared flux deficiency , the most striking example of which is lhs 1126 labeled in the right panel of figure [ fg : vmi : vmk ] .
as mentioned above , such stars have been interpreted as having mixed h / he / c atmospheres with @xmath18 @xcite . @xcite
have recently cast some doubt on the schmidt et al .
interpretation based of their own analysis of the chemical abundance equilibrium of a mixture of h / he / c / o and heavier elements .
however they have explored atmospheric parameters which do not match those inferred for c2h stars : hydrogen abundances about three orders of magnitude too small and effective temperatures much cooler ( @xmath121 k ) than the narrow range of @xmath122 6200 k where such stars are found .
the presence of trace amounts of hydrogen can be deduced from the presence of a broad and shallow h@xmath3 absorption feature in an otherwise helium - rich atmosphere due to the van der waals broadening of the hydrogen lines by neutral helium .
this is the case for ross 640 ( 1626@xmath68368 ) and l745@xmath6746a ( 0738@xmath67172 ) whose fits are displayed in figure [ fg : samplehe ] .
also shown in this figure is the fit to lhs 290 ( 1043@xmath67188 ) , a c@xmath5h star whose optical spectrum is shown in figure 30 of brl but for which we had not obtained any photometry at the time of writing the brl analysis ( see fig .
31 of brl for fits to additional c@xmath5h stars ) . in figure
[ fg : samplehe ] , the effective temperature and surface gravity of each object is derived from the fits to the energy distribution , while the hydrogen - to - helium abundance ratio is determined from fits to the h@xmath3 line profile ; only a limit to the abundance can be obtained in the case of lhs 290 .
the effective temperatures and abundances we derive for ross 640 and l745@xmath6746a are in excellent agreement with those determined by @xcite and references therein .
overluminous objects in the @xmath12 vs ( @xmath51@xmath78 ) color - magnitude diagram displayed in figure [ fg : mv : vmi ] are most likely unresolved binaries composed of two normal white dwarfs .
the masses inferred for these objects in a color - magnitude diagram are well below 0.4 @xmath6 ; note that two unresolved 0.6 @xmath6 white dwarfs would appear in such a diagram as a single @xmath123 @xmath6 white dwarf .
alternatively , it is also possible that the overluminous objects in figure [ fg : mv : vmi ] correspond to single low - mass ( @xmath124 @xmath6 ) white dwarf stars .
sample fits for five of those overluminous objects are displayed in figure [ fg : sampledoubleda ] . as mentioned above ,
all of these are of the da spectral type , and the surface gravities determined by fitting the solid angle are well below the average for normal da stars .
otherwise , the fits to the energy distributions do not reveal any additional peculiarity , and in particular , they are all consistent with single white dwarfs . if we were dealing with double degenerates with much different effective temperatures and luminosities , one could expect the combined energy distribution to be somewhat distorted .
the results of figure [ fg : sampledoubleda ] suggest that either these are single low - mass white dwarfs , or double degenerates with comparable atmospheric parameters .
the evidence that we are actually dealing with unresolved double degenerates stems from the fits to the h@xmath3 line profiles also shown in figure [ fg : sampledoubleda ] . in the case of a single low - mass star
, the fit to the h@xmath3 line profile would not show any discrepancy .
those shown here , however , are certainly at odds with the atmospheric parameters obtained from the energy distributions ( contrast the fits shown in figure [ fg : sampledoubleda ] with those shown in figure [ fg : sampleda ] ) . the only marginal case for which the theoretical line profile provides a satisfactory fit is l870@xmath672 ( 0135@xmath67052 ) , a well - known double degenerate system discovered by @xcite !
@xcite have actually determined that the effective temperatures of both components of the l870@xmath672 system differ by only 500 k ( @xmath125 and 7470 k ) , a result which explains why the combined h@xmath3 line profile can be reproduced by a single spectrum with an intermediate effective temperature . for the other stars in figure [ fg : sampledoubleda ] , the line profile fit discrepancies are much more obvious .
g21@xmath6715 ( 1824@xmath68040 ) is also a known double degenerate with a period of 6.266 days @xcite .
@xcite have reported a null result for g1@xmath6745 ( 0101@xmath68048 ) in their radial velocity survey , however .
nothing is known about the two remaining objects in figure [ fg : sampledoubleda ] .
our sample includes magnetic white dwarfs , several of which have been discovered and analyzed in the course of our survey ( see brl and references therein ) .
here we present additional fits to magnetic white dwarfs as well as magnetic candidates .
figure [ fg : figmag ] shows our best fits to the magnetic white dwarfs g128@xmath6772 ( 2329@xmath68267 ) and gd 175 ( 1503@xmath67070 ) , both of which exhibit zeeman splitting at h@xmath3 .
as discussed above , the magnetic nature of g128@xmath6772 has been reported by @xcite while gd 175 represents a new discovery . here
the effective temperatures and surface gravities have been taken from our best fits to the energy distributions .
the synthetic magnetic spectra are calculated using offset dipole models following the prescription outlined in @xcite .
both the value of the dipole field strength , @xmath126 , and the dipole offset , @xmath127 , are considered free parameters , while the viewing angle @xmath128 between the dipole axis and the line of sight is determined from a visual inspection of the fits . for g128@xmath6772 ,
our best fit is achieved with @xmath129 mg , @xmath130 , and @xmath131 , in good agreement with the values derived by @xcite , @xmath132 mg and @xmath133 , under the assumption of a centered dipole model .
note that the shape of the shifted zeeman components are particularly useful to determine the viewing angle , as shown for instance in figure 4 of @xcite .
gd 175 is an obvious case of an unresolved double degenerate binary containing a magnetic da component and a featureless dc component , similar to the g62@xmath6746 system analyzed by ( * ? ? ?
* see also the case of lhs 2273 shown in fig .
33 of brl ) . in such systems ,
the h@xmath3 line profile predicted under the assumption of a single magnetic da star is much stronger than the observed line profile , as shown in figure [ fg : figmag ] for gd 175 ( dotted line ) .
the simplest explanation for this discrepancy is that gd 175 is an unresolved double degenerate binary system composed of a magnetic da star whose emergent flux is diluted by the presence of a featureless dc companion .
the thick solid line in the bottom panel represents our best solution with the assumption that this dc white dwarf contributes equally to the total continuum flux near h@xmath3 .
we do not attempt here to derive the atmospheric parameters for the individual components of the system , as carried out for g62@xmath6746 by @xcite .
instead we note that if two identical but different chemical compositions white dwarfs are contributing to the total flux , the mass of the magnetic da component is near 1 @xmath6 , as opposed to @xmath134 @xmath6 when a single star is assumed .
cool da white dwarfs with much weaker magnetic fields will not show any zeeman splitting , at least not at the spectroscopic resolution used here .
a good example of such an object is the @xmath135 kg white dwarf lhs 1038 ( 0009@xmath68501 ) reported by @xcite and analyzed photometrically by brl ( see their fig .
the spectroscopic fit at h@xmath3 using non - magnetic models is reproduced in figure [ fg : samplemag ] .
for all objects in this figure , the atmospheric parameters are those obtained from the energy distributions . as mentioned in brl , the theoretical line profile for lhs 1038
is predicted somewhat too deep , as a result of the magnetic splitting of the h@xmath3 line core into its unresolved zeeman components .
a similar object has also been reported by brl ( see their fig .
30 ) , lhs 5064 ( 0257@xmath68080 ) , and its fit at h@xmath3 is reproduced in figure [ fg : samplemag ] as well .
brl made the suggestion that lhs 5064 was also a weakly magnetic ( @xmath136 kg ) da star .
recently , @xcite observed this object using the technique of zeeman spectropolarimetry and indeed detected a weak ( @xmath137 kg ) magnetic field at h@xmath3 .
three additional objects in our survey exhibit similar features lhs 2800 ( 1344@xmath68106 ) , g138@xmath6747 ( 1635@xmath68137 ) , and g156@xmath6764 ( 2253@xmath67081 ) , and the fits at h@xmath3 are also shown in figure [ fg : samplemag ] .
@xcite has reported a null result on all three of these objects , however .
it is always possible that an unfortunate orientation of the magnetic axis prevents the detection of circular polarization .
given the resemblance of these fits , it would be worth observing these stars at high spectral resolution to detect any weak zeeman splitting .
alternative explanations for these three peculiar fits to the h@xmath3 line profile include the presence of helium in their atmosphere .
experiments with mixed h / he models , not shown here , indicate that by the time the predicted depth of the h@xmath3 line _ core _ agrees with the observed profile , the line _ wings _ are predicted much too broad .
another possibility is the presence of an unresolved dc white dwarf which would dilute the absorption profile of the da star .
however , the trigonometric parallax measurements of these objects make them subluminous in an absolute magnitude - color diagram , implying larger than average masses , and inconsistent with these stars being unresolved degenerate binaries as they would then be superluminous .
finally , the remaining object displayed in figure [ fg : samplemag ] is lhs 3501 ( 1953@xmath67011 ) recently reported to be magnetic by ( * ? ? ? * and references therein ) , with a field strength of @xmath91 kg although the zeeman triplet at h@xmath3indicates a field strength of only @xmath99 kg .
the line profile is also found to be variable over a time - scale of a day or less .
our line fit shown here appears to be consistent with a non - magnetic da star , however .
it is possible that our spectroscopic observation was secured during a rotational phase when zeeman splitting was the least apparent .
figure [ fg : figg227 - 35 ] shows our best fit to the energy distribution of the magnetic white dwarf g227@xmath6735 ( 1829@xmath68547 ) .
this star has been analyzed in detail by @xcite who convincingly demonstrated that this object is a strongly magnetic @xmath138 mg _ hydrogen - rich _ white dwarf .
their spectrum shows a prominent feature in circular polarization with a weak corresponding absorption feature which corresponds to the 2@xmath139 - 13@xmath140 - 2 transition of h@xmath3 .
however , as shown in figure [ fg : figg227 - 35 ] , the observed energy distribution is reproduced much better with a pure helium atmosphere rather than a pure hydrogen atmosphere !
the most likely explanation in this case is that the h@xmath23 continuum opacity in pure hydrogen magnetic models is sufficently disrupted in the presence of such a high magnetic field that the corresponding energy distribution resembles instead that of a pure helium model . _
eso 439@xmath6726_.as discussed in [ sbsc : optir ] , with the recent result that
extremely cool , hydrogen atmosphere white dwarfs become bluer , there are now two possible interpretations for the location of the non - da star eso 439@xmath6726 ( 1136@xmath67286 ) in the @xmath12 vs ( @xmath51@xmath78 ) color - magnitude diagram displayed in figure [ fg : mv : vmi ] .
the first one is that the object is a massive ( @xmath141 @xmath6 ) , pure helium atmosphere white dwarf , i.e. the interpretation proposed by @xcite . the second one is that the object is a much cooler ( @xmath142 k ) , and thus older , lower mass white dwarf .
our best fit to the complete energy distribution using our pure helium model grid is displayed in figure [ fg : figeso439 - 26 ] .
our solution , @xmath84 k and @xmath143 , which corresponds to @xmath144 @xmath6 , is consistent with the location of eso 439@xmath6726 in the right panel of figure [ fg : mv : vmi ] ( i.e. the pure helium sequences ) .
also shown in figure [ fg : figeso439 - 26 ] is our best fit to the @xmath51 and @xmath78 measurements using our pure hydrogen model grid .
this time our solution , @xmath145 k and @xmath146 , which corresponds to @xmath147 @xmath6 , is consistent with the location of eso 439@xmath6726 in the middle panel of figure [ fg : mv : vmi ] ( i.e. the pure hydrogen sequences ) .
however , the predicted flux at @xmath81 is clearly inconsistent with the observed flux , and the pure hydrogen solution must be rejected .
this result stresses the importance of obtaining infrared photometry for such cool white dwarfs before drawing any conclusion about their temperature and especially their age . _
g107@xmath6770a / b_.g107@xmath6770a / b ( 0727@xmath68482a / b ) is a partially resolved pair of cool degenerate stars with an orbital period of 20.5 yr @xcite .
this system has long been thought to be composed of two dc white dwarfs , but our new combined spectrum shown in figure [ fg : plotspecda ] clearly shows a weak h@xmath3 absorption feature , a result which indicates that at least one of the two components of the system has some hydrogen in its atmosphere .
a visual inspection of the spectra of the coolest da stars in figure [ fg : plotspecda ] shows that the h@xmath3 profile of g107@xmath6770a / b is much broader and shallower than the other da spectra with similar line strengths .
this suggests that the system may be composed of a normal da star whose h@xmath3 absorption line profile is diluted by the contribution of a featureless white dwarf component . in the following we attempt to determine the atmospheric parameters of the individual components of the g107@xmath6770a / b system . the magnitude difference between both white dwarfs of the g107@xmath6770a
/ b system has been measured to be 0.3 mag at @xmath51 , @xmath81 , @xmath82 , and @xmath52 ; we assume the same difference at @xmath50 , @xmath77 , and @xmath78 .
we have obtained fits to the a and b components under the assumption that one star has a pure hydrogen atmosphere , while the other one has either a pure hydrogen or pure helium atmosphere ( for a total of three possible combinations ) .
then the theoretical combined spectrum is calculated as the sum of both interpolated spectra , properly weighted by the luminosity of each star .
our results indicate that acceptable fits to the energy distributions can always be achieved , no matter what atmospheric composition is assumed . on the other hand , the h@xmath3 line profile is always predicted much weaker and sharper than the observed line profile .
the dilution of the h@xmath3 line profile by the dc component does not alter significantly enough the _ shape _ of the combined line profile . our best solution shown in figure [ fg :
figg107 - 70 ] is obtained with two pure hydrogen atmosphere white dwarfs .
the discrepancy at h@xmath3 can not be attributed to an inadequacy with the models since we are able to fit single da stars in the same range of effective temperature ( see , e.g. , lhs 1801 in figure [ fg : sampleda ] ) .
we have also tried to restrict our analysis to only @xmath51 and the infrared where the magnitude difference has actually been measured , but without improvement .
experiments with mixed h / he atmospheres have not been successful either .
perhaps the h@xmath3 absorption feature of one of the da component , or both , is broadened by a weak magnetic field .
it is even possible that the g107@xmath6770 system is composed of three degenerate components ; after all , such systems are now known to exist @xcite .
we note finally that the combined mass of the pair with our adopted solution is @xmath148 @xmath6 , in excellent agreement with the astrometric value of @xmath149 @xmath6 derived by @xcite .
among the 152 white dwarfs listed in table 1 , 150 have been analyzed with the techniques described in this paper .
the stars omitted are l481@xmath6760 ( 1544@xmath67377 ) which has a bright star nearby that contaminates the photometric observations , and bpm 27606 ( 2154@xmath67512 ) whose c@xmath5 swan bands are so strong that the color indices are significantly affected and can not be modeled within the present theoretical framework .
as opposed to the brl analysis , all white dwarfs analyzed here have trigonometric parallax measurements which allow us to determine their radii , and hence their masses and surface gravities through the mass - radius relation .
the results of our analysis are reported in table 2 .
for each white dwarf we give the effective temperature , the surface gravity , the dominant atmospheric constituent , the stellar mass , the absolute visual magnitude , the total luminosity , and the white dwarf cooling age ( not including the main sequence lifetime ) .
the bolometric magnitudes can be derived from the simple expression , @xmath150 .
the numbers in parentheses in table 2 correspond to the uncertainties of each parameter .
the error of @xmath46 is derived directly from the fits to the energy distributions , while the errors of log @xmath151 and other quantities have been derived by propagating the error of the trigonometric parallax uncertainty .
we have also assumed single stars for all objects in table 2 , even for the suspected or known unresolved degenerate binaries , and a note mentioning the binary nature of these systems has been added . as discussed in [ sbsc : fittec ] , new c / o core evolutionary models with thin and thick hydrogen layers have been used for the helium- and hydrogen - dominated atmospheres , respectively .
we have assumed pure hydrogen and pure helium compositions for the hydrogen- and helium - rich white dwarfs , respectively , with the exception of the c@xmath5h stars , as well as ross 640 and l745@xmath6746a displayed in figure [ fg : samplehe ] , for which the helium abundances have been determined individually .
pure helium compositions were also assumed for the group of peculiar hot ( @xmath152 k ) non - da white dwarfs whose energy distributions are consistent with pure hydrogen models ( see brl ) ; a note in table 2 identifies these objects . for the magnetic white dwarf g227@xmath6735 shown in figure [ fg : figg227 - 35 ] , we adopt the pure helium solution since its energy distribution is better reproduced by the pure helium model rather than the pure hydrogen model , despite the fact it has a hydrogen - rich composition according to @xcite .
additional blue spectroscopic observations which cover the high balmer lines from h@xmath108 to h9 are also available to us for a subset of 24 cool da white dwarfs in our sample .
these observations can be used to obtain independent estimates of @xmath46 and @xmath55 using the spectroscopic technique ( see , e.g. , * ? ? ? * ) which consists of fitting all the hydrogen line profiles observed in each star simultaneously .
the value of @xmath55 can be converted into mass using an appropriate mass - radius relation for white dwarfs .
these independent determinations can be compared against the values derived here using optical / infrared photometry and trigonometric parallaxes .
the comparison of effective temperature and mass determinations for the 24 white dwarfs in common with our sample is displayed in figure [ fg : tspecvstphot ] .
seven objects discussed further below are numbered in this figure in order of increasing right ascension .
the effective temperatures agree well within the uncertainties of both techniques throughout the entire temperature range considered in this analysis , with the exception of the four stars labeled in the bottom panel of figure [ fg : tspecvstphot ] . of these ,
l587@xmath6777a ( 0326@xmath67273 ; labeled 2 ) and case 2 ( 1606@xmath68422 ; labeled 3 ) are both low - mass ( @xmath153 @xmath6 ) white dwarfs , and they are most likely unresolved degenerate binaries with individual components having much different effective temperatures , while g21@xmath6715 ( 1824@xmath68040 ; labeled 6 ) is a known double degenerate ( * ? ? ? * ; * ? ? ?
* see our fit in fig .
[ fg : sampledoubleda ] ) .
this comparison indicates that photometry combined with spectroscopy can be used to identify unresolved degenerate binary candidates . the last object , g128@xmath6772 ( 2329@xmath68267 ; labeled 7 ) , is a magnetic white dwarf with our best fit at h@xmath3 already shown in figure [ fg : figmag ] . in this case , the spectroscopic temperature is overestimated since the blue spectrum of this star has been fit with non - magnetic models , and magnetic line broadening can be compensated , to some extent , by an increase in effective temperature ( and @xmath55 , see below ) .
the agreement between the photometric masses and the spectroscopic masses for this sample of 24 da stars , shown in the top panel of figure [ fg : tspecvstphot ] , is not as good as that between the effective temperatures , even within the mass uncertainties . in general , the spectroscopic masses are larger than those obtained from trigonometric parallax measurements .
there are a few exceptions , however , and these are labeled in the top panel of figure [ fg : tspecvstphot ] .
first , g19@xmath6720 ( 1716@xmath68020 ; labeled 5 ) has a photometric mass of @xmath154 @xmath6 and a spectroscopic mass of @xmath155 @xmath6 .
an independent mass estimate for this white dwarf star comes from the gravitational redshift measurement which yields a value of @xmath156 @xmath6 @xcite , a result which suggests that the trigonometric parallax measurement is probably in error for this particular object .
second , g31@xmath6735 ( 0011@xmath68000 ; labeled 1 ) , has a fairly large mass uncertainty ( @xmath157 @xmath6 ) due to a @xmath158% error in the trigonometric parallax measurement , and the difference from the spectroscopic mass of @xmath159 @xmath6 is not significant . the objects whose spectroscopic masses are larger than the photometric masses include the three low - mass white dwarfs ( objects labeled 2 , 3 , and 6 ) already identified as binaries or binary candidates in the bottom panel of figure [ fg : tspecvstphot ] , and g128@xmath6772 ( 2329@xmath68267 ; labeled 7 ) the magnetic white dwarf discussed above . for the three binaries ,
the photometric masses are underestimated since the total flux received from the system is assumed to originate from a single object , while the spectroscopic masses are intermediate values of the masses of the individual components , as demonstrated by @xcite . if we assume instead that two identical white dwarfs contribute to the total flux received at earth , we obtain the photometric masses indicated by the three vectors in figure [ fg : tspecvstphot ] , and these agree perfectly with the spectroscopic masses for two of these objects , while the third object is probably composed of two white dwarfs with different luminosities . as discussed in
[ sbsc : hrich ] , there are da stars for which the agreement between the photometric and spectroscopic solutions is excellent , the best example of which is g74@xmath677 ( 0208@xmath68396 ) displayed in figure [ fg : sampleda ] .
why then are there several da white dwarfs whose spectroscopic masses are significantly larger than the photometric masses ?
what is the missing parameter , if any ?
@xcite has suggested that the convective mixing of a thin hydrogen atmosphere with the deeper and more massive helium envelope , which is believed to occur in hydrogen - rich white dwarfs cooler that @xmath160 k , could result in atmospheres which are only moderately enriched with helium rather than in helium - dominated atmospheres .
since helium is spectroscopically invisible at these temperatures , the mixed h / he white dwarfs would still appear as normal da stars .
* see also @xcite ) has also demonstrated that normal gravity , helium - enriched atmospheres are totally equivalent , _ from a spectroscopic point of view _
, to pure hydrogen atmospheres but with large surface gravities ( and hence masses ) .
hydrogen line profiles at a given temperature are affected by atmospheric pressure , which in turn is sensitive to changes in the surface gravity and in the helium abundance , and both effects can not be disentangled from spectroscopy alone .
in contrast , the photometric technique provides a direct measure of the star radius which is not very sensitive to the presence of helium .
hence , the mass discrepancies observed in figure [ fg : tspecvstphot ] could well be the result of helium enrichment in some of the white dwarfs in our sample .
we illustrate the effect of adding small amounts of helium to the atmospheres of cool da white dwarfs in figure [ fg : examplehe ] for lhs 3254 ( 1655@xmath68215 ) , the most discrepant case in figure [ fg : tspecvstphot ] ( labeled 4 in the top panel ) .
the top panels show our best photometric fit ( left panel ) and our best spectroscopic fit ( right panel ) using our pure hydrogen model grid .
even though the effective temperatures agree within @xmath161 k , the surface gravity determined from spectroscopy , @xmath162 ( or @xmath163 @xmath6 ) , is much larger than that obtained from the photometric solution , @xmath164 ( or @xmath165 @xmath6 ) . by allowing a small amount of helium in the model grid , in this case @xmath166 , we achieve the solutions displayed in the bottom panels of figure [ fg : examplehe ] .
first , the photometric fit has changed little , both qualitatively and quantitatively , a result which implies that the overall results of our analysis remain unaffected by this new parameter .
second , the spectroscopic fit has not changed qualitatively , but the surface gravity has decreased significantly , and most importantly , the photometric and spectroscopic @xmath55 values ( and hence masses ) are now in perfect agreement .
the effective temperatures are also in better agreement .
hence , our results reinforce the conclusions of @xcite that several but probably not all cool da stars have helium - enriched atmospheres , with helium abundances as high as @xmath167 .
we note that ( see the following section ) the spectroscopically determined masses for hot db stars can also be brought into better agreement with our photometrically determined values if very small amounts of hydrogen are added to the pure helium model grid .
at least trace amounts of both hydrogen and helium are therefore likely to exist in many cool white dwarf atmospheres .
the mass of all white dwarfs in our sample is shown as a function of effective temperature in figure [ fg : spec_evol ] . each object
is represented by a different symbol indicating its atmospheric composition and/or spectral type . filled and open symbols represent hydrogen - rich and helium - rich atmospheric compositions , respectively .
the more specific symbols are discussed in the next section .
a comparison of this figure with figure 35 of brl indicates that our new sample has increased in size considerably , especially above @xmath168 k where the brl sample contains only a few objects .
furthermore , it is possible here to determine stellar masses for all white dwarfs in our sample , in contrast with brl who could only determine masses for about 50% of their sample .
while massive white dwarfs are found with both hydrogen and helium atmospheres , low mass white dwarfs have mostly hydrogen - rich atmospheres , with the exceptions of the dq star g184@xmath6712 ( 1831@xmath68197 ) at @xmath169 k , and lp 31@xmath6740 ( 0324@xmath68738 ) at @xmath170 k. the mass of the former object could be as large as 0.56 @xmath6 due to the large parallax uncertainty , while the latter object lacks the crucial @xmath2 photometry which would allow us to determine its atmospheric constituent more precisely , and this object may well have a hydrogen - rich atmosphere .
consequently , it is probable that all low - mass white dwarfs possess hydrogen - rich atmospheres .
this conclusion is supported by the spectroscopic analyses of hotter ( @xmath171 k ) da and db stars @xcite which revealed the existence of a low mass tail in the mass distribution of da stars which is not observed in the db sample .
these results are illustrated in figure [ fg : mass_distr ] where we compare the mass distributions of the hydrogen- and helium - rich atmosphere white dwarfs in our sample ( top panel ) and in the hotter spectroscopic samples ( bottom panel ) ; note that the photometric and spectroscopic samples do not overlap in effective temperature .
since common envelope evolution is required to produce degenerates with mass @xmath172 @xmath6 the galaxy being too young to have produced them from single star evolution we must conclude that this particular evolutionary channel does not produce helium - rich atmosphere white dwarfs , presumably because the objects which go through this close - binary phase end up with hydrogen layers too massive to allow the da to db conversion near @xmath173 k @xcite , or any possible convective mixing below 12,000 k. the fraction of low - mass , hydrogen - rich atmosphere white dwarfs is larger in the parallax sample ( @xmath174% ) than in the spectroscopic sample ( @xmath175% ) .
this is due to the fact that in the parallax sample , unresolved double degenerate systems composed of two normal 0.6 @xmath6 white dwarfs , for example , will appear as a single overluminous low - mass @xmath123 @xmath6 star , while it will still appear as a single 0.6 @xmath6 object when analyzed spectroscopically .
there are significant differences at high masses as well where the trigonometric parallax sample contains more massive stars ( @xmath176 @xmath6 ) than the spectroscopic sample .
since the former sample is composed of older objects , it is possible that it contains several cases of mergers which would result in a single white dwarf with higher than average mass @xcite .
note that the most massive star in our sample with @xmath177 @xmath6 is eso 439@xmath6726 ( 1136@xmath67286 ) , discussed at length above .
the mean mass of the hydrogen - rich parallax subsample is @xmath178 @xmath6 with a dispersion of @xmath179 @xmath6 , while we find for the helium - rich subsample a mean mass of @xmath180 @xmath6 with a dispersion of @xmath181 @xmath6 ; the combined sample yields @xmath182 @xmath6 , @xmath179 @xmath6 .
the lower mean value of the hydrogen - rich set is entirely due to the existence of the low - mass tail component .
these numbers can be compared with the values derived for the spectroscopic analyses of hotter da stars , @xmath183 @xmath6 and @xmath184 @xmath6 @xcite , and of db stars , @xmath183 @xmath6 and @xmath185 @xmath6 @xcite . note that the dispersions of the spectroscopic mass distributions are considerably smaller due to the increased sensitivity to @xmath55 of the spectroscopic technique over the trigonometric parallax method . while the mean masses obtained for the cool hydrogen atmospheres and warmer da stars
agree well , those of the helium atmospheres and db stars differ by @xmath186 @xmath6 .
as discussed in brl , these results can be somewhat reconciled if small amounts of undetectable hydrogen is present in the cool helium - rich atmospheres .
quantitative results not shown here indicate that the masses derived from the pure helium models could be reduced by @xmath187 @xmath6 were they analyzed with a @xmath188 model grid .
it is worth mentioning that the spectroscopic analyses of da and db stars discussed above are not sensitive to the presence of unknown traces of hydrogen or helium .
indeed , da stars in these studies are restricted to a range of temperature above 15,000 k where the assumption of pure hydrogen atmospheres is most certainly valid due to the gravitational settling of helium @xcite .
similarly , db ( and dba ) stars have temperatures in excess of 12,000 k where the hydrogen abundance is well determined , or at least constrained @xcite .
hence , the spectroscopic mass distributions displayed in figure [ fg : mass_distr ] are not sensitive to unknown abundances of trace elements , unlike the results for cool ( @xmath0 k ) da stars discussed in [ sbsc : compspec ] and shown in figure [ fg : tspecvstphot ] , in which the presence of spectroscopically invisible helium could overestimate the spectroscopic masses . out of the 150 objects plotted in figure
[ fg : spec_evol ] , 96 have a hydrogen - rich atmosphere while 54 have a helium - rich atmosphere .
hence 36% of all white dwarfs in our sample have helium - rich atmospheres , a ratio which is significantly larger than the canonical value of 20% for hotter white dwarfs @xcite .
the ratio determined here , however , is not constant over the temperature range considered in this analysis .
figure [ fg : h_vs_he ] shows the number of hydrogen - rich and helium - rich stars as a function of effective temperature in 1000 k bins using the results presented in table 2 ; the corresponding hydrogen- to helium - rich ratio is also shown .
the @xmath45000 - 6000 k bin shows a marked increase in the number of white dwarfs with hydrogen - rich atmospheres , and a marked drop in the number with helium - rich atmospheres .
this is the non - da gap first identified by brl . our new results , with a much larger sample and with no overt spectroscopic classification bias , confirms the reality of this gap .
note also that there is no particular concentration of non - da stars below @xmath15 k , nor is there any particular concentration of da stars around 6000 k , contrary to the claim of @xcite .
quite the opposite , hydrogen- and helium - rich white dwarfs in figure [ fg : spec_evol ] are more or less homogeneously distributed in effective temperature , with the exception of the non - da gap itself .
we thus reaffirm our earlier conclusion that some as yet unidentified physical mechanism is responsible for turning helium - rich atmosphere white dwarfs into hydrogen - dominated atmospheres near @xmath189 k. the fact that the cooling timescales of these white dwarfs may depend on their atmospheric composition , as observed by @xcite , has nothing to do with the conclusions reached here .
indeed , if the existence of this gap is the result of the spectral evolution of helium - rich into hydrogen - rich white dwarfs , the only important factor is the mass - dependency of the mechanism through which this transition occurs .
if this mechanism still unknown is only weakly mass - dependent , then helium - rich white dwarfs will turn into da stars near 6000 k , no matter what their cooling age might be . an alternative explanation for the non - da gap could be that the blue end of the gap at 6000 k represents the end of the cooling sequence for helium - rich white dwarfs , and that all non - da stars observed below the red end of the gap are the result of convectively mixed da stars at 5000 k. however , the cooling age of helium - rich white dwarfs near 6000 k is only around 4 gyr , and since much older da stars exist , older and cooler helium - rich white dwarfs should exist as well .
the spectroscopic behavior seen in our sample around the gap is also of interest .
to summarize , figure [ fg : spec_evol ] shows that : hydrogen - rich white dwarfs are seen at all effective temperatures ; dq stars are seen at @xmath190 k ; dz at @xmath190 k and 45005500 k ; helium - rich da s at 80009000 k ; c@xmath5h stars at 55006500 k ; and the peculiar group whose featureless energy distributions are better fit with pure hydrogen models at 60007500 k. at temperatures just above and near the non - da gap , non - da white dwarfs that are neither dq nor dz belong to the peculiar group of white dwarfs whose featureless energy distributions are better fit with pure hydrogen models .
one exception directly above the gap is the dc star g125@xmath673 ( 1917@xmath68386 ) with @xmath191 k. according to @xcite , this object could have barely visible , heavily broadened c@xmath5 bands .
their spectrum which is available to us reveals three weak and broad absorption features which actually match those of c@xmath5h stars , and it is possible that this object belongs to this spectral type , although better spectroscopic observations are required before this result is confirmed .
near 7500 k , there are three additional dc stars .
the coolest object is stein 2051b ( 0426@xmath68588 ) with @xmath192 k , a true dc star according to @xcite who reported featureless optical as well as ultraviolet spectroscopic observations .
the second dc star is g195@xmath6719 ( 0912@xmath68536 , @xmath193 k ) , a polarized @xmath194 mg magnetic white dwarf @xcite , and the last one is lhs 2333 ( 1055@xmath67072 , @xmath195 k ) whose fit has already been discussed in [ sbsc : herich ] and displayed in figure [ fg : samplenonda ] .
another helium - rich object with only a slightly higher effective temperature is gd 356 ( 1639@xmath68537 , @xmath196 k ) , a dae white dwarf whose spectrum exhibits a hydrogen zeeman triplet in emission @xcite . most non - da white dwarfs within the non - da gap are c@xmath5h stars .
g240@xmath6772 ( 1748@xmath68708 ) is considered in this spectral class as well , following the interpretation of brl that the observed 15% deep yellow sag near 5000 is the result of c@xmath5h molecular absorption features in the @xmath197 mg magnetic field which characterizes this star .
near the red edge of the gap are lhs 2710 ( 1313@xmath67198 ) and eso 292@xmath6743 ( 2345@xmath67447 ) which have already been identified by brl as being peculiar in that the energy distributions can not be matched by our models ( see their figure 37 and the accompanying discussion ) .
most importantly , _ our new larger sample has not added any new non - da star within the gap_. during the course of our observational survey , every time a previously classified non - da star was found inside the gap , subsequent high signal - to - noise spectroscopy would systematically reveal the presence of h@xmath3 .
note that the new magnetic candidate c@xmath5h star lhs 2229 recently reported by @xcite has an effective temperature of @xmath198 k when analyzed with our fitting technique ( assuming @xmath65 ) .
if this interpretation is correct , c@xmath5h stars may exist well below the red edge of the non - da gap .
the results shown in figure [ fg : spec_evol ] reveal that dq and c@xmath5h stars do not overlap in effective temperature .
indeed , there are no dq stars observed below @xmath199 k , while c@xmath5h stars are found only below this temperature .
this suggests that dq stars probably turn into c@xmath5h stars , and that the existence of the non - da gap and the nature of the c@xmath5h white dwarfs are somehow related . since c@xmath5h stars have been interpreted by @xcite as the result of mixed h / he / c compositions , it is the additional presence of hydrogen in a supposedly helium and carbon mixture that would turn dq stars into c@xmath5h stars when they cool off below 6400 k. the origin of the hydrogen source remains a mystery , however , since the exotic model proposed by brl has been discarded by @xcite based on detailed model simulations .
we just note that the same source of hydrogen , once identified , could be also responsible for turning the remaining non - da stars above @xmath200 k ( dz and peculiar ) into da stars below this temperature .
peculiar white dwarfs whose energy distributions are better reproduced with hydrogen models despite the absence of h@xmath3 are found over a range of effective temperature of 62007600 k , i.e. all above the non - da gap .
the hottest of these stars is lp 434@xmath6797 ( 1154@xmath68186 ) at @xmath201 k. interestingly , immediately above this temperature are the two helium - rich da stars in our sample , l745@xmath6746a ( 0738@xmath67172 ) at @xmath202 k , and ross 640 ( 1626@xmath68368 ) at @xmath203 k. given that there are no other such helium - rich da stars known , it is tempting to suggest that an evolutionary link exists between these objects and the peculiar white dwarfs . above 8000 k , most non - da stars are of the dq type ,
although some are reported to be featureless in the literature .
the number of stars in this temperature range is still too small , however , to draw meaningful conclusions about any spectral evolution from one type to another .
the reappearance of non - da stars below the gap has been interpreted by brl as the result of the convective mixing of the hydrogen atmosphere with the deeper helium envelope , turning da stars into non - da stars .
the calculations reported by brl showed indeed that envelope models which include realistic model atmospheres as boundary conditions could reduce the mixing temperature limit from the canonical value of @xmath20 k down to 5000 k , which occurs for a hydrogen layer mass of @xmath204 .
obviously , some da stars have thicker hydrogen layers since they did not mix , in particular the ( apparent or true ) low - mass white dwarfs which are either unresolved binaries , or the result of common envelope evolution .
we note that the coolest non - da stars in our sample have hydrogen - rich atmospheres despite the lack of any detectable hydrogen absorption features .
while h@xmath3 can be detected almost to the end of the disk cooling sequence , it will certainly not be detected in any cooler white dwarfs discovered in the galactic halo .
our sample contains a higher ratio of non - da to da stars than is seen in the temperature range 12,000 - 30,000 k , where the ratio is around 25% @xcite .
this increase can certainly be interpreted as the result of convective mixing over a still unknown temperature range . since
the 10,00011,000 k bin in figure [ fg : h_vs_he ] already shows a ratio near unity , and since the reappearance of non - da stars in large number below the non - da gap has also been interpreted as the result of convective mixing , we can conclude that the convective mixing mechanism operate from @xmath205 k , all the way down to the coolest possible mixing temperature near 5000 k. this implies that da stars must possess a large range of possible hydrogen layer thicknesses , including very thick hydrogen layers above @xmath206 . out of the 13 magnetic white dwarfs in our sample
8 with hydrogen - rich and 5 with helium - rich atmospheres , 12 are located in a narrow range of effective temperature between @xmath207 and 8000 k ; the exception is gw@xmath6870 8247 ( 1900@xmath68705 ) at @xmath208 k. it is possible that the lower boundary represents only a selection effect since zeeman splitting would make the detection of h@xmath3 more difficult in cool magnetic da stars than in non - magnetic white dwarfs .
however , the spectrum of the coolest magnetic star in our sample , lhs 1734 ( 0503@xmath67174 ) , still exhibits strong h@xmath3 absorption features ( see the spectrum in fig . 5 of brl ) , and the detection of even weaker features is well within the detection limit of our spectroscopic observations .
furthermore , while there are no magnetic white dwarfs in the three dozen or so objects included in our sample between @xmath209 and 12,000 k , a few magnetic stars in this particular range of effective temperature are known to exist @xcite .
all the features of the chemical distribution shown in figure [ fg : spec_evol ] can not be accounted for by convective mixing alone , however , or by any simple theoretical model .
hence we are left with very few possible explanations , despite the abundance of observational clues unveiled in brl and in this analysis .
given these results , it is worth repeating and developing an alternative explanation proposed by brl who questioned the stability of convectively mixed hydrogen and helium convection zones .
consider a 0.6 @xmath6 , pure helium atmosphere white dwarf which accretes @xmath210 @xmath6 of hydrogen by the time it reaches 6000 k. this value is totally plausible if one allows a low accretion rate from the interstellar medium of @xmath211 @xmath6 yr@xmath212 over a period of @xmath213 gyr or so .
since the mass of the helium convection zone is almost constant below @xmath214 ( @xmath215 ) , the resulting abundance of @xmath49 would make the star appear as a non - da white dwarf at 6000 k , h@xmath3 being wiped out by van der waals broadening with neutral helium .
it is reasonable to wonder whether such a mixed hydrogen and helium convection zone would remain convectively stable as the results shown in figure 40 of brl indicate that a da star with a hydrogen layer mass of @xmath210 @xmath6 or slightly larger , would not have mixed in the first place , at least not at @xmath216 k. it is possible that convectively mixed hydrogen and helium atmospheres return to a well separated hydrogen - dominated atmosphere with a helium - rich envelope .
note that the chemical separation needs to be complete only in the photospheric layers for the atmosphere to appear hydrogen - rich .
such `` born again ''
da stars can not retain a da signature for long , however . as the star cools from @xmath216
k down to 5000 k , the bottom of the hydrogen atmosphere further plunges from @xmath217 to @xmath218 ( depending on the mass of the star , see fig . 4 of brl ) , leading to another episode of convective mixing , and supporting the trends seen in figure [ fg : spec_evol ] .
if the separation of hydrogen and helium layers does occur , we can speculate about what the energy distribution of these stars might look like at the moment just before the separation occurs .
obviously , the hydrogen and helium abundance profiles throughout the atmosphere will be very inhomogeneous , even over the disk surface .
perhaps this would explain why brl ( see also * ? ? ?
* ) were unable to fit the energy distributions of those peculiar white dwarfs above 6000 k with pure hydrogen , pure helium , or _
homogeneously mixed _ hydrogen and helium compositions .
models with inhomogeneous abundance profiles could hold the key to these peculiar stars .
we note finally that if the total mass of hydrogen is small , the convectively mixed hydrogen and helium layer is stable
. this can certainly be the case for the c@xmath5h stars with mixed atmospheres of hydrogen , helium , and carbon , as well as ross 640 and l745@xmath6746a , which are simply too hot and thus too young to have accreted enough hydrogen ( these objects have @xmath37 ) .
the isochrones for our c / o core evolutionary models with thin and thick hydrogen layers are shown in figures [ fg : isochrone_co_thin ] and [ fg : isochrone_co_thick ] , respectively . in these plots , we no longer distinguish between objects with hydrogen- and helium - rich atmospheres since the atmospheric composition of these stars may have changed several times over their lifetime , and as such , it is not a trivial task to associate a cooling age to a single object , as discussed in the introduction . also shown
are the corresponding isochrones but with the main sequence lifetime added to the white dwarf cooling age ; here we simply assume @xcite @xmath219 gyr and @xmath220 $ ] .
as can be seen from these results , white dwarfs with @xmath221 @xmath6 can not have c / o cores , and yet have been formed from single star evolution within the lifetime of this galaxy . as discussed above , some of these low - mass objects must be either unresolved double degenerates , or single white dwarfs with helium cores . in the former case ,
the stellar masses inferred from these figures are underestimated especially if the unresolved components have comparable luminosities , and the corresponding cooling ages derived here become meaningless . while some of these low - mass white dwarfs may still be interpreted as having c / o cores when the masses are corrected to allow for two objects in the system , the lowest mass objects in our sample would still yield `` corrected '' masses too low to be c / o - core degenerates .
the second possibility corresponds to single ( or binary ) helium - core degenerates whose core mass was truncated by case b mass transfer before helium ignition was reached .
such low - mass white dwarfs would be best described with helium - core evolutionary models . however , since it is impossible to distinguish between the two possibilities discussed above for most of the low - mass stars in our sample , we refrain from interpreting these objects any further . in this temperature regime ,
low - mass ( i.e. large - radius ) white dwarfs evolve faster than their more massive counterparts , with the exception of the most massive objects ( @xmath176 @xmath6 ) in which the onset of crystallization reduces the cooling timescales considerably .
these begin to cool off much more rapidly than the lower mass stars to produce the observed parabola - shaped isochrones . with decreasing effective temperature
, crystallization gradually occurs in lower mass models , and the turning point of these parabolas moves slowly towards lower masses .
consequently , the inferred stellar ages are strongly mass - dependent , a result which stresses the importance of determining reliable masses through precise trigonometric parallax measurements
. the best example is provided by eso 439@xmath6726 ( 1136@xmath67286 ) , the most massive object in our sample , which could have an inferred cooling age anywhere between 5 and 7.5 gyr if its mass was not known ( assuming thin hydrogen layer models ) .
the effect of using different thicknesses for the outer hydrogen layer varies as a function of the mass of the star as well .
for instance , the coolest white dwarfs near 0.6 @xmath6 have cooling ages of roughly 8 gyr with the thin hydrogen layer models , and 1 gyr _ older _ with the thick models . out of the three oldest objects in our sample , two have hydrogen - rich atmospheres , the coolest of which is er 8 ( 1310@xmath67472 ) with @xmath222 k and mass near 0.6 @xmath6 .
if we make the assumption that this star has a thick hydrogen envelope , the corresponding white dwarf cooling age is @xmath223 gyr ( @xmath224 gyr ) . the oldest object in figure [ fg : isochrone_co_thick ] , lp 701@xmath6729 ( 2251@xmath67070 ) with @xmath225 k , has a helium - rich atmosphere and it thus can not have had a thick outer hydrogen layer , otherwise it would have remained hydrogen - rich .
its more probable age derived from figure [ fg : isochrone_co_thin ] is actually much shorter , @xmath226 gyr ( @xmath227 gyr ) .
there are two objects in our sample cooler than er 8 with hydrogen - rich atmospheres with apparently low masses of @xmath228 @xmath6 , lhs 239 ( 0747@xmath68073b , @xmath72 k ) and lp 131@xmath6766 ( 1247@xmath68550 , 4050 k ) .
if these objects are in fact an unresolved pair of normal mass white dwarfs then the total age of these systems is around 10 gyr .
a detailed photometric and spectroscopic analysis of 152 cool white dwarfs with trigonometric parallax measurements was presented .
optical and infrared photometry were combined and compared with an improved grid of model atmosphere calculations to determine the effective temperatures and the solid angles for all objects , which combined with the measured parallaxes yielded the stellar radii .
the latter were then converted into stellar masses through the mass - radius relations obtained from new evolutionary models which best characterize cool white dwarfs .
these have carbon / oxygen cores , with both thin @xmath229 and thick @xmath230 hydrogen layers .
this parallax sample more than doubles the number of cool white dwarfs with known masses compared to the sample presented in brl .
furthermore , this new sample is more representative of the population of cool white dwarf stars as it contains , presumably , no bias in favor of any particular spectral type , as opposed to the brl sample which favored non - da stars .
one of the main results of our analysis is the confirmation of the existence of a range in effective temperature between roughly 5000 and 6000 k in which almost all white dwarfs have hydrogen - rich atmospheres .
the only non - da stars in this range of effective temperature are the so - called c@xmath5h stars which are believed to represent the cooler stage of the dq sequence , found only above the gap .
clearly , the existence of this gap holds the key to our understanding of the chemical evolution of cool white dwarfs .
unfortunately , our understanding of the physical mechanisms that could alter the atmospheric composition of cool white dwarfs as they evolve are at best sketchy . after the `` phase transition '' model proposed in brl
was convincingly discarded by @xcite , very few alternatives exist .
it has become clear from our analysis , however , that convection plays an important role , but that our understanding of what to expect from the result of convective mixing needs to be revised . it is generally believed that when the bottom of the thin hydrogen convection zone deepens and reaches the underlying and more massive convective helium envelope , convective mixing will occur , and that the resulting atmospheric composition will simply be governed by the mass ratio of the hydrogen and helium convection zones .
such arguments lead to predicted abundances after convective mixing of @xmath231 , @xmath232 , @xmath233 , and @xmath234 , for mixing temperatures of @xmath235 , 10,000 , 8500 , and 6500 k , respectively , the mixing temperature depending on the exact thickness of the hydrogen layer .
our analysis has demonstrated that such objects do not exist , at least in our sample .
even the objects which possess mixed hydrogen and helium atmospheres , and in particular ross 640 and l745@xmath6746a , have abundances that do not fit this simple convective mixing picture . on the other hand , the definite increase of the ratio of non - da to da stars in the temperature range considered in this analysis with respect to that of hotter white dwarfs can not easily be explained in terms other than convective mixing .
also , convective mixing has been called upon to explain the reappearance of non - da stars below the non - da gap near 5000 k. a different picture is now starting to emerge from our findings , and those of other analyses , in which convective mixing _ processes _ may lead to predicted chemical abundances that differ from the standard model .
first , our comparison of atmospheric parameters determined from photometry and spectroscopy reveals that a mass discrepancy exists in several da stars , and that this discrepancy can be resolved if we allow some amount of helium to be present in their atmospheres .
the required abundances close to @xmath236 are clearly at odds with the standard convective mixing model predictions , as noted also by @xcite .
note that helium is spectroscopically invisible in the temperature regime considered here and can not be detected directly .
it is possible to conceive that helium is being brought to the surface not by convective mixing , but perhaps by convective overshooting of the helium convection zone into the hydrogen layer .
the bottom of the hydrogen convection zone would then carry some of this helium to the surface , in a very similar fashion to the way carbon is being brought to the photosphere of dq stars , with the exception that ordinary diffusion is responsible in this case for polluting the bottom of the helium layer with carbon .
a variance of the convective model that could account for the existence of the non - da gap has also been proposed .
the question arises when one considers a helium - rich atmosphere on top of which indefinite amounts of hydrogen are accumulated , presumably from interstellar accretion .
it is clear that the hydrogen abundance ratio will gradually increase until some point where the homogeneously mixed hydrogen and helium convection zone can no longer be sustained . a stable chemical separation between the hydrogen and the helium layers
could then occur , leading to the predominance of da stars below @xmath216 k. the implication of this proposed model is that cool white dwarfs must accrete reasonable amounts of hydrogen from the interstellar medium for this mechanism to work .
also , it was proposed that the chemical separation could occur in a non - homogeneous fashion , leading to chemically inhomogeneous atmospheres , similar to the hotter dab stars ( see * ? ? ?
* ; * ? ? ?
such inhomogeneous atmospheres could perhaps account for the peculiar objects identified in our sample whose energy distributions share both characteristics of hydrogen and helium atmospheres .
given the complexity of the chemical evolution of cool white dwarf atmospheres revealed in our analysis , it is not clear what evolutionary models should be used to derive cooling ages .
thin and thick hydrogen layer models yield cooling ages which differ by 2 gyr in some cases .
non - da stars are probably better characterized by thin hydrogen models ; even if they were once da stars , they probably had thin hydrogen layers anyway in order to eventually mix . but cool da stars may have thin or thick hydrogen layers .
thick hydrogen layer white dwarfs remain da forever , but the fraction of such stars is still unknown .
the problem becomes even more complex when one attempts to determine the age of the galactic disk using the luminosity function of cool white dwarfs .
the theoretical luminosity functions should in fact reflect the actual population of cool white dwarfs , and should be a combination of thin and thick hydrogen layer models .
the problems are greatly simplified for hotter da , db , or do stars . from the results presented here , the oldest white dwarfs in our sample have a minimum stellar age of 8.5 gyr , but
are more likely characterized by models with thick hydrogen layers which yield an age around 9.5 gyr .
we feel at this point that we have pushed both the observational and theoretical aspects of this cool white dwarf study to their limits , and that more work now needs to be done on the theoretical front to explain the various features discovered in this analysis .
several avenues of investigation have been proposed , most of which are related to our poor understanding of convective mixing .
we wish to acknowledge g. fontaine for calculating the cooling models used in our analysis , p. brassard for making his evolutionary code available to us , and g. d. schmidt for his spectropolarimetric observations of the magnetic white dwarf candidates .
we are deeply grateful to the various observatories for awarding us so many nights for this project and to all the staff at the observatories who were extremely helpful .
we also wish to thank the referee for pointing us out the results of jorgensen et al .
of which we were unaware .
this work was supported in part by the nserc canada and by the fund fcar ( qubec ) .
mtr received partial support from fondecyt grant 1980659 and catedra presidencial en ciencias .
p2.3cmlrrrllllclllcl 0000@xmath67345&lhs 1008 & 75.7 & 9.0&0.0 & 15.02 & @xmath680.44 & @xmath680.30 & @xmath680.57 & 1&14.17 & @xmath680.15 & @xmath680.15 & 1&1 , 2 + 0009@xmath68501&lhs 1038 & 90.6 & 3.7&8.91 & 14.36 & @xmath680.45 & @xmath680.28 & @xmath680.59 & 1&13.41 & @xmath680.15 & @xmath680.05 & 1&1 , 2 , 3 , 4 + 0011@xmath68000&g31@xmath6735 & 32.9 & 4.8&34.8 & 15.35 & @xmath680.20 & @xmath680.08 & @xmath680.18 & 1&15.21 & @xmath680.08 & @xmath680.01 & 1&2 + 0011@xmath67134&lhs 1044 & 51.3 & 3.8&7.15 & 15.89 & @xmath680.63 & @xmath680.33 & @xmath680.67 & 1&14.85 & @xmath680.23 & @xmath680.10 & 1&1 , 2 , 4 , 5 + 0029@xmath67032&lhs 1093 & 42.6 & 1.0&0.0 & 17.32 & @xmath681.12 & @xmath680.61 & @xmath681.14 & 1&15.56 & @xmath680.19 & @xmath680.02 & 4&1 , 2 + + 0033@xmath68016&g1@xmath677 & 30.4 & 4.0&44.2 & 15.61 & @xmath680.18 & @xmath680.02 & @xmath680.08 & 1&15.63 & @xmath680.03 & @xmath670.03 & 1&2 + 0038@xmath67226&lhs 1126 & 101.2&10.4&0.0 & 14.50 & @xmath680.70 & @xmath680.42 & @xmath680.79 & 2&13.32 & @xmath670.15 & @xmath670.24 & 2&1 , 2 + 0038@xmath68555&g218@xmath678 & 43.4 & 2.0&0.0 & 14.05 & @xmath680.05 & @xmath680.04 & @xmath680.12 & 1&13.97 : & @xmath670.11 : & @xmath670.05 : & 1&2 , 6 + 0046@xmath68051&vma 2 & 232.5 & 1.9&0.0 & 12.39 & @xmath680.52 & @xmath680.26 & @xmath680.49 & 2&11.69 & @xmath680.08 & @xmath680.09 & 1&1 , 2 , 5 + 0101@xmath68048&g1@xmath6745 & 46.9 & 3.8&18.0 & 14.00 & @xmath680.26 & @xmath680.17 & @xmath680.34 & 1&13.51 & @xmath680.12 & @xmath680.01 & 1&2 + + 0115@xmath68159&lhs 1227 & 64.9 & 3.0&0.0 & 13.85 & @xmath680.10 & @xmath680.11 & @xmath680.20 & 1&13.72 & @xmath680.00 & @xmath670.02 & 1&1 , 2 + 0117@xmath67145&lhs 1233 & 34.8 & 1.2&1.05 & 16.96 & @xmath680.89 & @xmath680.51 & @xmath680.98 & 1&15.47 & @xmath680.30 & @xmath680.09 & 1&2 , 7 + 0121@xmath68401&g133@xmath678 & 32.1 & 5.5&4.17 & 17.12 & @xmath680.79 & @xmath680.45 & @xmath680.90 & 1&15.64 & @xmath680.21 & @xmath680.02 & 2&2 , 7 + 0126@xmath68101&g2@xmath6740 & 28.4 & 3.1&15.9 & 14.40 & @xmath680.24 & @xmath680.12 & @xmath680.28 & 1&14.05 & @xmath680.13 & @xmath670.02 & 1&2 , 6 + 0135@xmath67052&l870@xmath672 & 81.0 & 2.8&11.5 & 12.86 & @xmath680.33 & @xmath680.23 & @xmath680.48 & 1&12.12 & @xmath680.18 & @xmath680.02 & 1&1 , 2 + + 0142@xmath68312&g72@xmath6731 & 28.2 & 4.3&28.0 & 14.78 & @xmath680.22 & @xmath680.11 & @xmath680.26 & 1&14.38 & @xmath680.05 & @xmath670.05 & 1&2 + 0208@xmath68396&g74@xmath677 & 59.8 & 3.5&12.2 & 14.51 & @xmath680.33 & @xmath680.25 & @xmath680.47 & 1&13.80 & @xmath680.15 & @xmath680.02 & 1&1 , 2 , 5 + 0213@xmath68427&lhs 153 & 50.2 & 4.1&3.64 & 16.22 & @xmath680.73 & @xmath680.45 & @xmath680.85 & 1&14.98 & @xmath680.25 : & @xmath680.21 : & 2&2 , 5 , 6 + 0222@xmath68648&lhs 1405 & 31.4 & 0.8&0.0 & 18.29 & @xmath681.30 & @xmath680.71 & @xmath681.34 & 1&16.35 & @xmath680.29 & @xmath680.08 & 2&1 , 2 + 0230@xmath67144&lhs 1415 & 64.0 & 3.9&3.20 & 15.77 & @xmath680.69 & @xmath680.43 & @xmath680.84 & 1&14.43 & @xmath680.26 & @xmath680.06 & 1&1 , 2 + + 0243@xmath67026&lhs 1442 & 47.1 & 5.0&10.2 & 15.54 & @xmath680.40 & @xmath680.26 & @xmath680.53 & 2&14.71 & @xmath680.22 & @xmath670.02 & 2&1 , 2 + 0245@xmath68541&lhs 1446 & 96.6 & 3.1&0.73 & 15.36 & @xmath680.92 & @xmath680.52 & @xmath681.00 & 1&13.89 & @xmath680.23 & @xmath680.06 & 2&1 , 2 , 6 + 0257@xmath68080&lhs 5064 & 35.9 & 3.5&7.09 & 15.90 & @xmath680.43 & @xmath680.28 & @xmath680.58 & 1&15.01 & @xmath680.17 & @xmath680.04 & 1&1 , 2 , 4 + 0324@xmath68738&lp 31@xmath6740 & 25.0 & 3.8&0.0 & 17.68 & @xmath681.27 & @xmath680.77 & @xmath681.39 & 1&&&&0&2 + 0326@xmath67273&l587@xmath6777a & 57.6&13.6&25.3 & 14.00 & & & & 1&13.27 & @xmath680.15 & @xmath680.04 & 1&2 , 8 + + 0341@xmath68182&wolf 219 & 52.6 & 3.0&0.0 & 15.19 & @xmath680.33 & @xmath680.28 & @xmath680.54 & 1&14.56 & @xmath680.21 & @xmath670.05 & 1&1 , 2 , 5 + 0357@xmath68081&lhs 1617 & 56.1 & 3.7&5.97 & 15.92 & @xmath680.70 & @xmath680.41 & @xmath680.83 & 1&14.59 & @xmath680.26 & @xmath680.07 & 2&1 , 2 + 0426@xmath68588&stein 2051b & 180.6 & 0.8&0.0 & 12.43 & @xmath680.30 & @xmath680.26 & @xmath680.57 & 1&11.84 & @xmath680.11 & @xmath680.05 & 1&2 , 5 , 6 + 0433@xmath68270&g39@xmath6727 & 60.2 & 2.9&4.02 & 15.81 & @xmath680.67 & @xmath680.41 & @xmath680.80 & 1&14.61 & @xmath680.29 & @xmath680.10 & 1&1 , 2 + 0435@xmath67088&l879@xmath6714 & 105.2 & 2.6&0.0 & 13.75 & @xmath680.33 & @xmath680.32 & @xmath680.57 & 2&13.00 & @xmath680.15 & @xmath680.06 & 1&1 , 2 , 5 + + 0440@xmath68510&g175@xmath6746 & 21.6 & 5.2&19.0 & 15.95 & @xmath680.24 & @xmath680.13 & @xmath680.26 & 1&15.60 & @xmath680.10 & @xmath670.03 & 1&2 + 0503@xmath67174&lhs 1734 & 45.6 & 4.0&3.82 & 16.01 & @xmath680.72 & @xmath680.44 & @xmath680.90 & 1&14.55 & @xmath680.22 & @xmath680.10 & 2&1 , 2 , 4 + 0518@xmath68333&g86@xmath67b1b & 15.3 & 2.2&20.1 & 16.06 & @xmath680.22 & @xmath680.19 & @xmath680.41 & 1&15.52 & @xmath680.19 & @xmath670.01 & 1&2 + 0548@xmath67001&g99@xmath6737 & 90.3 & 2.8&0.0 & 14.56 & @xmath680.52 & @xmath680.34 & @xmath680.61 & 1&13.73 & @xmath680.10 & @xmath680.00 & 2&1 , 2 , 4 + 0551@xmath68468&lhs 1801 & 32.3 & 1.5&2.18 & 17.23 & @xmath680.84 & @xmath680.47 & @xmath680.94 & 1&15.84 & @xmath680.29 & @xmath680.02 & 2&2 , 7 + + 0552@xmath67041&lp 658@xmath672 & 155.0 & 2.1&0.0 & 14.47 & @xmath681.01 & @xmath680.50 & @xmath680.98 & 2&13.02 & @xmath680.12 & @xmath680.08 & 2&1 , 2 , 5 + 0553@xmath68053&g99@xmath6747 & 125.0 & 3.6&8.23 & 14.16 & @xmath680.62 & @xmath680.38 & @xmath680.75 & 2&12.96 & @xmath680.19 & @xmath680.11 & 2&1 , 2 , 4 , 5 + 0618@xmath68067&lhs 1838 & 44.2 & 4.2&5.52 & 16.41 & @xmath680.56 & @xmath680.36 & @xmath680.74 & 2&15.29 & @xmath680.24 & @xmath680.05 & 2&1 , 2 + 0644@xmath68025&g108@xmath6726 & 54.2 & 5.5&15.3 & 15.71 & @xmath680.35 & @xmath680.23 & @xmath680.46 & 2&15.00 & @xmath680.15 & @xmath670.08 & 2&1 , 2 + 0648@xmath68641&lp 58@xmath6753 & 30.0 & 5.2&6.24 & 16.65 & @xmath680.56 & @xmath680.36 & @xmath680.74 & 1&15.46 & @xmath680.27 & @xmath680.07 & 1&2 + + 0651@xmath67479&eso 207@xmath6721 & 16.1 & 1.6&0.0 & 19.54 & @xmath681.21 & @xmath680.65 & @xmath681.24 & 3&17.70 & @xmath680.23 & @xmath680.02 & 2&1 , 9 + 0654@xmath68027&g108@xmath6742 & 26.0 & 3.7&0.0 & 16.17 & @xmath680.12 & @xmath680.15 & @xmath680.20 & 1&15.98 & @xmath680.00 & @xmath680.00 & 1&2 + 0657@xmath68320&lhs 1889 & 53.5 & 0.9&1.60 & 16.62 & @xmath681.02 & @xmath680.53 & @xmath681.01 & 1&14.99 & @xmath680.22 & @xmath680.08 & 5&1 , 2 + 0659@xmath67064&lhs 1892 & 81.0&24.2&8.31 & 15.43 & @xmath680.43 & @xmath680.30 & @xmath680.60 & 1&14.58 & @xmath680.29 & @xmath680.05 & 1&2 , 5 + 0706@xmath68377&g87@xmath6729 & 41.2 & 2.4&0.0 & 15.64 & @xmath680.30 & @xmath680.26 & @xmath680.48 & 1&15.00 & @xmath680.12 & @xmath680.06 & 1&1 , 2 + + 0727@xmath68482a&g107@xmath6770a & 90.0 & 1.0&1.41 & 15.26 & @xmath680.98 & @xmath680.53 & @xmath681.02 & 1&13.66 & @xmath680.24 & @xmath680.09 & 3&1 , 2 , 5 , 7 + 0727@xmath68482b&g107@xmath6770b & 90.0 & 1.0&1.41 & 15.56 & @xmath680.98 & @xmath680.53 & @xmath681.02 & 1&13.96 & @xmath680.24 & @xmath680.09 & 3&1 , 2 , 5 , 7 + 0738@xmath67172&l745@xmath6746a & 112.4 & 2.7&4.88 & 13.06 & @xmath680.24 & @xmath680.18 & @xmath680.34 & 1&12.65 & @xmath680.04 & @xmath680.09 & 3&2 , 5 + 0743@xmath67340&vb 3 & 57.5 & 3.1&0.0 & 16.60 & @xmath681.21 & @xmath680.64 & @xmath681.21 & 2&14.85 & @xmath680.14 & @xmath680.16 & 2&1 , 10 + 0747@xmath68073a&lhs 240 & 54.7 & 0.7&0.41 & 16.63 & @xmath681.06 & @xmath680.58 & @xmath681.08 & 2&14.96 & @xmath680.23 & @xmath680.01 & 2&1 , 2 , 5 + + 0747@xmath68073b&lhs 239 & 54.7 & 0.7&0.0 & 16.96 & @xmath681.21 & @xmath680.65 & @xmath681.26 & 2&15.05 & @xmath680.15 & @xmath680.04 & 2&1 , 2 , 5 + 0751@xmath68578&g193@xmath6778 & 28.4 & 3.8&0.0 & 15.08 & @xmath680.09 & @xmath680.08 & @xmath680.17 & 1&14.94 & @xmath680.00 & @xmath670.02 & 1&2 + 0752@xmath67676&bpm 4729 & 141.2 & 8.4&4.70 & 13.95 & @xmath680.65 & @xmath680.37 & @xmath680.75 & 1&12.79 & @xmath680.27 & @xmath680.09 & 2&1 , 2 + 0752@xmath68365&g90@xmath6728 & 29.5 & 4.3&13.2 & 16.09 & @xmath680.29 & @xmath680.18 & @xmath680.40 & 1&15.50 & @xmath680.15 & @xmath680.00 & 1&2 + 0802@xmath68386&lp 257@xmath6728 & 24.0 & 3.0&0.0 & 15.56 & @xmath680.04 & @xmath680.05 & @xmath680.08 & 1&15.60 & @xmath680.02 & @xmath670.06 & 1&2 + + 0816@xmath68387&g111@xmath6771 & 25.2 & 2.5&13.1 & 16.51 & @xmath680.30 & @xmath680.22 & @xmath680.44 & 1&15.87 & @xmath680.15 & @xmath670.01 & 1&2 + 0827@xmath68328&lhs 2022 & 44.9 & 3.8&13.4 & 15.73 & @xmath680.32 & @xmath680.22 & @xmath680.47 & 1&15.01 & @xmath680.16 & @xmath680.01 & 1&2 + 0839@xmath67327&l532@xmath6781 & 112.7 & 9.7&26.7 & 11.90 & @xmath680.25 & & & 0&11.59 & @xmath680.04 & @xmath680.00 & 1&2 , 8 + 0856@xmath68331&g47@xmath6718 & 48.8 & 3.4&0.0 & 15.16 & @xmath680.03 & @xmath680.13 & @xmath680.19 & 1&15.12 & @xmath680.03 & @xmath670.02 & 1&1 , 2 + 0912@xmath68536&g195@xmath6719 & 97.3 & 1.9&0.0 & 13.84 & @xmath680.35 & @xmath680.20 & @xmath680.33 & 2&13.22 & @xmath680.07 & @xmath680.06 & 1&1 , 2 , 4 , 5 + + 0913@xmath68442&g116@xmath6716 & 34.6 & 4.0&22.0 & 15.37 & @xmath680.22 & @xmath680.15 & @xmath680.30 & 1&14.96 & @xmath680.12 & @xmath670.03 & 1&2 + 0930@xmath68294&g117@xmath6725 & 31.2 & 4.6&21.2 & 15.98 & @xmath680.24 & @xmath680.15 & @xmath680.32 & 1&15.51 & @xmath680.11 & @xmath670.05 & 1&2 + 0946@xmath68534&g195@xmath6742 & 43.5 & 3.5&0.0 & 15.18 & @xmath680.17 & @xmath680.13 & @xmath680.28 & 1&14.90 & @xmath680.03 & @xmath670.01 & 2&1 , 2 + 0955@xmath68247&g49@xmath6733 & 40.9 & 4.5&20.1 & 15.06 & @xmath680.24 & @xmath680.12 & @xmath680.32 & 1&14.66 & @xmath680.07 & @xmath670.06 & 1&2 + 1012@xmath68083&g43@xmath6738 & 34.5 & 3.9&8.68 & 16.03 & @xmath680.41 & @xmath680.27 & @xmath680.57 & 1&15.20 & @xmath680.21 & @xmath680.05 & 1&2 + + 1019@xmath68637&lp 62@xmath67147 & 61.2 & 3.6&11.5 & 14.70 & @xmath680.38 & @xmath680.25 & @xmath680.53 & 1&13.83 & @xmath680.20 & @xmath670.02 & 1&2 + 1039@xmath68145&g44@xmath6732 & 22.2 & 3.5&0.0 & 16.51 & @xmath680.28 & @xmath680.22 & @xmath680.41 & 1&15.93 & @xmath680.07 & @xmath680.08 & 2&1 , 2 + 1043@xmath67188&lhs 290 & 56.9 & 6.5&0.0 & 15.52 & @xmath680.57 & @xmath680.49 & & 1&14.62 & @xmath680.21 & @xmath680.05 & 1&2 , 5 + 1055@xmath67072&lhs 2333 & 82.3 & 3.5&0.0 & 14.33 & @xmath680.30 & @xmath680.20 & @xmath680.42 & 1&13.81 & @xmath680.10 & @xmath680.02 & 1&2 , 5 + 1108@xmath68207&lhs 2364 & 38.3 & 2.7&0.0 & 17.70 & @xmath680.97 & @xmath680.53 & @xmath681.07 & 2&15.91 & @xmath680.22 & @xmath680.07 & 2&1 , 2 , 5 + + 1115@xmath67029&lhs 2392 & 26.3 & 5.1&0.0 & 15.34 & @xmath680.08 & @xmath680.09 & @xmath680.17 & 1&15.23 & @xmath670.04 : & @xmath670.02 : & 1&2 + 1121@xmath68216&ross 627 & 74.4 & 2.8&14.0 & 14.21 & @xmath680.31 & @xmath680.20 & @xmath680.45 & 1&13.58 & @xmath680.18 & @xmath680.00 & 1&2 , 5 + 1124@xmath67296&eso 439@xmath6780 & 16.4 & 1.7&29.3 & 15.02 & & @xmath680.07 & @xmath680.26 & 1&14.90 & @xmath680.02 & @xmath680.08 & 1&1 , 9 + 1136@xmath67286&eso 439@xmath6726 & 24.5 & 1.0&0.0 & 20.52 & @xmath681.03 & @xmath680.64 & @xmath681.14 & 3&18.64 & & & 1&1 , 11 + 1142@xmath67645&lhs 43 & 218.3 & 6.7&0.0 & 11.49 & @xmath680.17 & @xmath680.16 & @xmath680.29 & 1&11.19 & @xmath680.07 & @xmath680.03 & 1&2 + + 1147@xmath68255&lp 375@xmath6751 & 19.5 & 2.9&43.9 & 15.66 & @xmath680.15 & @xmath680.06 & @xmath680.13 & 1&15.53 & @xmath680.04 & @xmath670.02 & 1&2 + 1154@xmath68186&lp 434@xmath6797 & 33.5 & 3.5&0.0 & 15.61 & @xmath680.26 & @xmath680.18 & @xmath680.37 & 1&15.15 & @xmath680.12 & @xmath680.01 & 1&2 + 1208@xmath68576&lhs 2522 & 48.9 & 4.6&4.87 & 15.78 & @xmath680.57 & @xmath680.37 & @xmath680.74 & 1&14.64 & @xmath680.25 & @xmath680.07 & 1&2 + 1215@xmath68323&g148@xmath67b4b & 32.2 & 3.6&0.0 & 17.00 & @xmath680.33 & @xmath680.23 & @xmath680.60 & 2&16.39 & @xmath680.15 & @xmath670.13 : & 1&2 + 1236@xmath67495<t 4816 & 61.0 & 9.4&45.5 & 13.80 & @xmath680.18 & @xmath670.02 & @xmath670.03 & 1&13.92 & @xmath680.02 & @xmath670.08 & 1&2 + + 1244@xmath68149&g61@xmath6717 & 15.7 & 4.2&39.6 & 15.89 & @xmath680.17 & @xmath680.03 & @xmath680.06 & 1&15.84 & @xmath670.02 & @xmath680.02 & 1&2 + 1247@xmath68550&lp 131@xmath6766 & 39.6 & 0.7&0.0 & 17.79 & @xmath681.44 & @xmath680.76 & @xmath681.45 & 1&15.72 & @xmath680.05 & @xmath680.04 & 2&1 , 2 , 5 + 1257@xmath68037&lhs 2661 & 60.3 & 3.8&5.07 & 15.84 & @xmath680.66 & @xmath680.38 & @xmath680.76 & 2&14.56 & @xmath680.23 & @xmath680.08 & 2&1 , 2 , 5 + 1257@xmath68278&g149@xmath6728 & 28.9 & 4.1&20.4 & 15.41 & @xmath680.23 & @xmath680.15 & @xmath680.28 & 1&14.99 & @xmath680.08 & @xmath680.00 : & 1&2 + 1300@xmath68263&lhs 2673 & 28.4 & 3.3&0.0 & 18.77 & @xmath681.24 & @xmath680.68 & @xmath681.28 & 2&16.89 & @xmath680.18 & @xmath680.01 & 2&5 , 12 + + 1310@xmath67472&er 8 & 66.5 & 2.4&0.0 & 17.13 & @xmath681.39 & @xmath680.72 & @xmath681.41 & 2&15.21 & @xmath680.10 & @xmath680.08 & 1&1 , 2 + 1313@xmath67198&lhs 2710 & 40.6 & 1.8&0.0 & 17.14 & @xmath680.95 & @xmath680.43 & @xmath680.82 & 2&15.87 & @xmath680.17 & @xmath680.14 & 4&1 , 9 + 1325@xmath68581&g199@xmath6771 & 27.0 & 5.5&9.74 & 16.70 & @xmath680.40 & @xmath680.28 & @xmath680.54 & 1&15.82 & @xmath680.14 & @xmath680.03 & 1&2 + 1328@xmath68307&g165@xmath677 & 33.4 & 5.3&0.0 & 16.03 & @xmath680.70 & @xmath680.29 & @xmath680.43 & 1&15.50 & @xmath680.14 & @xmath680.02 & 1&2 + 1334@xmath68039&wolf 489 & 121.4 & 3.4&1.19 & 14.63 & @xmath680.95 & @xmath680.51 & @xmath681.01 & 1&13.06 & @xmath680.26 & @xmath680.10 & 2&1 , 2 , 5 + + 1344@xmath68106&lhs 2800 & 49.9 & 3.6&11.6 & 15.12 & @xmath680.38 & @xmath680.22 & @xmath680.48 & 1&14.38 & @xmath680.18 & @xmath680.01 & 1&2 , 5 + 1345@xmath68238&lp 380@xmath675 & 82.9 & 2.2&0.46 & 15.71 & @xmath681.15 & @xmath680.59 & @xmath681.13 & 2&13.92 & @xmath680.25 & @xmath680.08 & 2&1 , 2 , 5 + 1418@xmath67088&g124@xmath6726 & 24.6 & 3.8&17.9 & 15.38 & @xmath680.28 & @xmath680.17 & @xmath680.38 & 1&14.81 & @xmath680.12 & @xmath680.00 & 1&2 + 1444@xmath67174&lhs 378 & 69.0 & 4.0&0.0 & 16.44 & @xmath681.03 & @xmath680.49 & @xmath681.01 & 1&14.94 & @xmath680.15 & @xmath680.11 & 2&1 , 2 , 5 + 1455@xmath68298&lhs 3007 & 28.9 & 4.1&13.0 & 15.60 & @xmath680.31 & @xmath680.22 & @xmath680.42 & 1&14.86 & @xmath680.13 & @xmath680.01 & 1&2 + + 1503@xmath67070&gd 175 & 38.5 & 5.5&9.08 & 15.88 & @xmath680.40 & @xmath680.28 & @xmath680.55 & 1&15.07 & @xmath680.14 & @xmath680.02 & 1&2 , 4 + 1544@xmath67377&l481@xmath6760 & 74.6&10.1&33.8 & 12.97 : & @xmath680.18 : & @xmath680.22 : & @xmath680.60 : & 1&&&&0&2 , 13 + 1606@xmath68422&case 2 & 22.2 & 3.5&48.8 & 13.82 & @xmath680.11 & @xmath670.03 & @xmath670.06 & 1&13.92 & @xmath680.00 & @xmath670.09 & 1&2 + 1609@xmath68135&lhs 3163 & 54.5 & 4.7&25.9 & 15.11 & @xmath680.20 & @xmath680.10 & @xmath680.24 & 1&14.77 & @xmath680.01 & @xmath680.01 & 1&2 + 1625@xmath68093&g138@xmath6731 & 42.8 & 3.7&10.8 & 16.14 & @xmath680.34 & @xmath680.27 & @xmath680.52 & 3&15.34 & @xmath680.22 & @xmath680.06 & 3&1 , 2 + + 1626@xmath68368&ross 640 & 62.7 & 2.0&8.99 & 13.83 & @xmath680.19 & @xmath680.08 & @xmath680.17 & 2&13.58 & @xmath680.01 & @xmath670.01 & 1&2 + 1633@xmath68433&g180@xmath6763 & 66.2 & 3.0&9.11 & 14.84 & @xmath680.43 & @xmath680.27 & @xmath680.56 & 1&13.95 & @xmath680.19 & @xmath680.03 & 1&2 + 1633@xmath68572&g225@xmath6768 & 69.2 & 2.5&0.0 & 14.99 & @xmath680.50 & @xmath680.31 & @xmath680.61 & 1&14.03 & @xmath680.06 & @xmath670.02 & 2&1 , 2 , 5 + 1635@xmath68137&g138@xmath6747 & 25.5 & 4.6&11.1 & 16.94 & @xmath680.38 & @xmath680.27 & @xmath680.47 & 2&16.11 & @xmath680.15 & @xmath670.02 : & 1&2 , 7 + 1637@xmath68335&g180@xmath6765 & 35.0 & 3.2&34.3 & 14.65 & @xmath680.18 & @xmath680.06 & @xmath680.14 & 1&14.56 & @xmath680.06 & @xmath670.04 & 1&2 + + 1639@xmath68537&gd 356 & 47.4 & 3.5&-4.85 & 15.05 & @xmath680.29 & @xmath680.21 & @xmath680.39 & 2&14.54 & @xmath680.08 & @xmath680.04 & 1&2 , 4 + 1655@xmath68215&lhs 3254 & 43.0 & 3.1&26.4 & 14.13 & @xmath680.21 & @xmath680.10 & @xmath680.21 & 1&13.89 & @xmath680.09 & @xmath670.05 & 1&2 + 1656@xmath67062&lp 686@xmath6732 & 30.7 & 0.9&4.21 & 17.14 & @xmath680.79 & @xmath680.34 & @xmath680.72 & 2&15.89 & @xmath680.25 & @xmath680.08 & 1&1 , 9 + 1705@xmath68030&g139@xmath6713 & 57.0 & 5.4&0.0 & 15.20 & @xmath680.43 & @xmath680.24 & @xmath680.46 & 1&14.62 & @xmath680.12 & @xmath680.02 & 1&2 + 1716@xmath68020&g19@xmath6720 & 28.1 & 2.6&60.2 & 14.36 & @xmath680.12 & @xmath670.04 & @xmath670.04 & 1&14.68 & @xmath680.03 & @xmath670.06 : & 1&2 + + 1733@xmath67544&l270@xmath67137 & 45.6 & 0.8&7.35 & 15.70 & @xmath680.46 & @xmath680.29 & @xmath680.53 & 2&14.89 & @xmath680.34 & @xmath680.09 & 1&1 , 9 + 1736@xmath68052&g140@xmath672 & 23.4 & 3.9&38.8 & 15.92 & @xmath680.23 & @xmath680.12 & @xmath680.26 & 1&15.62 & @xmath680.06 & @xmath680.07 & 1&2 + 1748@xmath68708&g240@xmath6772 & 164.7 & 2.4&0.0 & 14.13 & @xmath680.48 & @xmath680.53 & @xmath681.05 & 1&12.77 & @xmath680.07 & @xmath680.20 & 1&1 , 2 , 4 , 5 + 1756@xmath68827&lhs 56 & 63.9 & 2.9&12.1 & 14.34 & @xmath680.35 & @xmath680.22 & @xmath680.47 & 2&&&&0&2 , 5 + 1811@xmath68327a&g206@xmath6717 & 19.5 & 4.3&10.7 & 16.40 & @xmath680.30 & @xmath680.20 & @xmath680.41 & 1&15.71 & @xmath680.15 & @xmath680.02 & 1&2 + + 1811@xmath68327b&g206@xmath6718 & 19.5 & 4.3&6.82 & 17.07 & @xmath680.44 & @xmath680.28 & @xmath680.59 & 1&16.08 & @xmath680.14 & @xmath680.12 : & 1&2 , 6 + 1818@xmath68126&g141@xmath672 & 24.5 & 5.5&5.08 & 16.09 & @xmath680.51 & @xmath680.33 & @xmath680.67 & 1&15.07 & @xmath680.17 & @xmath680.03 & 1&1 , 2 + 1820@xmath68609&g227@xmath6728 & 78.2 & 4.1&0.82 & 15.69 & @xmath680.98 & @xmath680.54 & @xmath681.05 & 1&13.96 & @xmath680.23 & @xmath680.08 & 2&2 , 7 + 1824@xmath68040&g21@xmath6715 & 18.2 & 2.3&33.8 & 13.89 & @xmath680.09 & @xmath670.05 & @xmath670.07 & 1&14.07 & @xmath670.07 & @xmath680.00 & 1&2 , 6 + 1826@xmath67045&g21@xmath6716 & 34.9 & 3.8&23.8 & 14.58 & @xmath680.21 & @xmath680.11 & @xmath680.26 & 1&14.44 : & @xmath680.00 : & @xmath680.05 : & 1&2 , 6 + + 1829@xmath68547&g227@xmath6735 & 66.8 & 5.6&0.0 & 15.57 & @xmath680.49 & @xmath680.29 & @xmath680.60 & 1&14.76 & @xmath680.15 & @xmath680.11 & 1&2 , 4 + 1831@xmath68197&g184@xmath6712 & 17.9 & 4.8&0.0 & 16.41 & @xmath680.29 & @xmath680.23 & @xmath680.37 & 3&15.93 & @xmath680.11 & @xmath680.01 & 1&2 + 1840@xmath68042&gd 215 & 40.2 & 3.4&24.3 & 14.79 & @xmath680.22 & @xmath680.13 & @xmath680.27 & 1&14.53 : & @xmath680.07 : & @xmath670.04 : & 1&2 + 1855@xmath68338&g207@xmath679 & 30.5 & 4.4&54.8 & 14.64 & @xmath680.17 & @xmath680.01 & @xmath670.02 & 1&14.74 & @xmath680.02 & @xmath670.05 & 1&2 + 1900@xmath68705&gw+70 8247 & 77.0 & 2.3&0.0 & 13.25 & @xmath680.06 & @xmath680.01 & @xmath680.02 & 1&&&&0&2 , 4 + + 1917@xmath68386&g125@xmath673 & 85.5 & 3.4&0.0 & 14.61 & @xmath680.45 & @xmath680.30 & @xmath680.57 & 1&13.77 & @xmath680.08 & @xmath680.10 & 1&1 , 2 + 1953@xmath67011&lhs 3501 & 87.8 & 2.9&13.8 & 13.69 & @xmath680.28 & @xmath680.19 & @xmath680.38 & 2&13.12 & @xmath680.10 & @xmath680.00 & 1&2 , 4 + 2002@xmath67110&lhs 483 & 57.7 & 0.8&0.0 & 16.95 & @xmath681.16 & @xmath680.59 & @xmath681.09 & 2&15.32 & @xmath680.21 & @xmath680.02 & 2&1 , 2 , 5 , 6 + 2011@xmath68065&g24@xmath679 & 44.7 & 1.9&0.0 & 15.78 & @xmath680.38 & @xmath680.27 & @xmath680.53 & 2&14.94 & @xmath680.15 & @xmath680.04 & 2&1 , 2 + 2048@xmath68263&g187@xmath678 & 49.8 & 3.4&1.80 & 15.63 & @xmath680.92 & @xmath680.52 & @xmath681.00 & 1&14.12 & @xmath680.29 & @xmath680.04 & 2&1 , 2 + + 2054@xmath67050&vb 11 & 64.6 & 5.1&0.0 & 16.69 & @xmath681.20 & @xmath680.66 & @xmath681.32 & 1&14.82 & @xmath680.21 & @xmath680.07 & 1&1 , 2 , 5 + 2059@xmath68190&g144@xmath6751 & 26.1 & 4.4&7.91 & 16.36 & @xmath680.43 & @xmath680.26 & @xmath680.54 & 1&15.52 & @xmath680.16 & @xmath680.02 & 2&1 , 2 + 2059@xmath68247&g187@xmath6716 & 36.1 & 4.2&7.75 & 16.57 & @xmath680.59 & @xmath680.31 & @xmath680.64 & 1&15.45 & @xmath680.16 & @xmath680.03 & 1&2 , 7 + 2059@xmath68316&g187@xmath6715 & 29.0 & 3.5&0.0 & 15.04 & @xmath680.11 & @xmath680.07 & @xmath680.14 & 1&14.94 & @xmath670.03 & @xmath670.01 & 1&2 + 2105@xmath67820&l24@xmath6752 & 58.6 & 8.8&35.5 & 13.61 & @xmath680.21 & @xmath680.05 & @xmath680.14 & 1&13.52 & @xmath670.01 & @xmath670.05 & 1&2 + + 2107@xmath67216&lhs 3636 & 42.2 & 1.5&6.45 & 16.80 & @xmath680.59 & @xmath680.35 & @xmath680.77 & 1&15.63 & @xmath680.18 : & @xmath680.05 : & 2&1 , 9 + 2111@xmath68261&g187@xmath6732 & 31.4 & 3.7&18.2 & 14.68 & @xmath680.24 & @xmath680.16 & @xmath680.32 & 1&14.15 & @xmath680.07 & @xmath670.01 & 1&2 , 6 + 2136@xmath68229&g126@xmath6718 & 23.8 & 3.1&31.9 & 15.25 & @xmath680.17 & @xmath680.06 & @xmath680.14 & 1&15.04 & @xmath680.08 & @xmath670.13 & 1&2 , 6 + 2140@xmath68207&lhs 3703 & 79.9 & 3.2&0.0 & 13.24 & @xmath680.13 & @xmath680.14 & @xmath680.26 & 1&12.95 & @xmath680.02 & @xmath670.02 & 1&2 , 6 + 2154@xmath67512&bpm 27606 & 68.5&10.6&0.0 & 14.74 & @xmath680.19 & @xmath680.44 & @xmath680.61 & 1&13.47 & @xmath680.00 & @xmath680.18 & 1&1 , 2 , 13 + + 2207@xmath68142&g18@xmath6734 & 39.2 & 4.4&12.8 & 15.60 & @xmath680.29 & @xmath680.20 & @xmath680.42 & 1&14.99 & @xmath680.18 & @xmath670.03 & 1&2 , 6 + 2246@xmath68223&g67@xmath6723 & 52.5 & 4.1&49.2 & 14.39 & @xmath680.17 & @xmath680.05 & @xmath680.14 & 1&14.28 & @xmath670.03 & @xmath670.06 & 1&2 + 2248@xmath68293&g128@xmath677 & 47.8 & 4.2&4.70 & 15.54 & @xmath680.66 & @xmath680.40 & @xmath680.79 & 1&14.24 & @xmath680.23 & @xmath680.07 & 2&1 , 2 , 5 + 2251@xmath67070&lp 701@xmath6729 & 123.7 & 4.3&0.0 & 15.71 & @xmath681.84 & @xmath680.61 & @xmath681.15 & 2&13.86 & @xmath680.23 & @xmath680.16 & 2&1 , 2 , 5 + 2253@xmath67081&g156@xmath6764 & 36.7 & 5.3&8.45 & 16.48 & @xmath680.40 & @xmath680.27 & @xmath680.56 & 1&15.59 & @xmath680.12 & @xmath680.11 : & 2&2 + + 2311@xmath67068&g157@xmath6734 & 39.8 & 4.7&0.0 & 15.40 & @xmath680.23 & @xmath680.18 & @xmath680.35 & 1&14.98 & @xmath680.05 & @xmath680.03 & 2&2 , 6 + 2312@xmath67024&lhs 3917 & 37.5 & 5.9&0.0 & 16.31 & @xmath680.51 & @xmath680.26 & @xmath680.49 & 1&15.70 & @xmath680.17 & @xmath670.05 & 1&1 , 2 + 2316@xmath67064&lhs 542 & 32.2 & 3.7&0.0 & 18.15 & @xmath681.08 & @xmath680.62 & @xmath681.16 & 1&16.38 & @xmath680.24 & @xmath680.04 & 2&5 , 12 + 2329@xmath68267&g128@xmath6772 & 25.9 & 4.7&32.4 & 15.33 & @xmath680.27 & @xmath680.06 & @xmath680.17 & 2&15.13 & @xmath680.10 & @xmath670.15 & 1&2 + 2345@xmath67447&eso 292@xmath6743 & 38.1 & 2.2&0.0 & 17.87 & @xmath680.79 & @xmath680.42 & @xmath680.84 & 1&16.66 & @xmath680.07 & @xmath680.26 & 2&1 , 9 + + 2347@xmath68292&lhs 4019 & 46.5 & 4.1&4.12 & 15.76 & @xmath680.59 & @xmath680.35 & @xmath680.72 & 1&14.59 & @xmath680.24 & @xmath680.11 & 1&1 , 2 , 6 + 2352@xmath68401&g171@xmath6727 & 38.7 & 5.6&0.0 & 14.94 & @xmath680.19 & @xmath680.19 & @xmath680.32 & 1&14.57 & @xmath680.05 & @xmath680.02 & 1&1 , 2 + p2.3cmrlcllllc 0000@xmath67345 & 6240 ( 140 ) & 8.31 ( 0.16 ) & he&0.77 ( 0.11 ) & 14.41 ( 0.26 ) & @xmath673.85 ( 0.11 ) & 4.18 ( 0.89 ) & 1 + 0009@xmath68501 & 6540 ( 150 ) & 8.23 ( 0.06 ) & h & 0.74 ( 0.04 ) & 14.15 ( 0.09 ) & @xmath673.71 ( 0.03 ) & 2.97 ( 0.38 ) & 2 + 0011@xmath68000 & 9610 ( 260 ) & 8.40 ( 0.19 ) & h & 0.85 ( 0.12 ) & 12.94 ( 0.32 ) & @xmath673.15 ( 0.12 ) & 1.29 ( 0.52 ) & + 0011@xmath67134 & 6010 ( 120 ) & 8.20 ( 0.11 ) & h & 0.71 ( 0.07 ) & 14.44 ( 0.16 ) & @xmath673.84 ( 0.07 ) & 3.65 ( 0.79 ) & 2 + 0029@xmath67032 & 4770 ( 50 ) & 8.07 ( 0.04 ) & he&0.61 ( 0.02 ) & 15.47 ( 0.05 ) & @xmath674.17 ( 0.02 ) & 7.07 ( 0.18 ) & + + 0033@xmath68016&10700 ( 350 ) & 8.66 ( 0.16 ) & h & 1.02 ( 0.09 ) & 13.02 ( 0.29 ) & @xmath673.14 ( 0.12 ) & 1.67 ( 0.45 ) & + 0038@xmath67226 & 5400 ( 170 ) & 7.91 ( 0.17 ) & he&0.52 ( 0.10 ) & 14.53 ( 0.22 ) & @xmath673.87 ( 0.09 ) & 3.51 ( 1.21 ) & 3 + 0038@xmath68555&10900 ( 630 ) & 8.14 ( 0.07 ) & he&0.67 ( 0.04 ) & 12.24 ( 0.10 ) & @xmath672.77 ( 0.04 ) & 0.62 ( 0.05 ) & 4 + 0046@xmath68051 & 6770 ( 200 ) & 8.40 ( 0.01 ) & he&0.83 ( 0.01 ) & 14.22 ( 0.02 ) & @xmath673.77 ( 0.01 ) & 3.67 ( 0.04 ) & 5 + 0101@xmath68048 & 8080 ( 200 ) & 7.55 ( 0.14 ) & h & 0.37 ( 0.05 ) & 12.36 ( 0.18 ) & @xmath672.96 ( 0.07 ) & 0.63 ( 0.07 ) & 6 + + 0115@xmath68159 & 9800 ( 360 ) & 8.38 ( 0.06 ) & he&0.82 ( 0.04 ) & 12.91 ( 0.10 ) & @xmath673.10 ( 0.04 ) & 1.17 ( 0.16 ) & 4 + 0117@xmath67145 & 5150 ( 110 ) & 7.75 ( 0.06 ) & h & 0.44 ( 0.03 ) & 14.67 ( 0.07 ) & @xmath673.86 ( 0.03 ) & 2.63 ( 0.27 ) & + 0121@xmath68401 & 5340 ( 120 ) & 7.90 ( 0.29 ) & h & 0.52 ( 0.16 ) & 14.65 ( 0.38 ) & @xmath673.88 ( 0.15 ) & 2.93 ( 1.87 ) & + 0126@xmath68101 & 8500 ( 200 ) & 7.20 ( 0.17 ) & h & 0.25 ( 0.04 ) & 11.67 ( 0.24 ) & @xmath672.68 ( 0.10 ) & 0.43 ( 0.06 ) & 6 + 0135@xmath67052 & 7140 ( 160 ) & 7.19 ( 0.06 ) & h & 0.25 ( 0.01 ) & 12.40 ( 0.07 ) & @xmath673.00 ( 0.03 ) & 0.66 ( 0.03 ) & 6 + + 0142@xmath68312 & 8660 ( 210 ) & 7.51 ( 0.25 ) & h & 0.35 ( 0.10 ) & 12.03 ( 0.33 ) & @xmath672.82 ( 0.13 ) & 0.52 ( 0.11 ) & 6 + 0208@xmath68396 & 7310 ( 180 ) & 8.01 ( 0.09 ) & h & 0.60 ( 0.05 ) & 13.39 ( 0.13 ) & @xmath673.38 ( 0.05 ) & 1.38 ( 0.17 ) & + 0213@xmath68427 & 5600 ( 160 ) & 8.12 ( 0.12 ) & h & 0.66 ( 0.08 ) & 14.72 ( 0.18 ) & @xmath673.91 ( 0.07 ) & 3.89 ( 1.09 ) & + 0222@xmath68648 & 4520 ( 40 ) & 8.00 ( 0.05 ) & he&0.57 ( 0.03 ) & 15.77 ( 0.06 ) & @xmath674.23 ( 0.02 ) & 7.08 ( 0.46 ) & + 0230@xmath67144 & 5480 ( 120 ) & 8.11 ( 0.09 ) & h & 0.65 ( 0.06 ) & 14.80 ( 0.13 ) & @xmath673.95 ( 0.05 ) & 4.28 ( 0.94 ) & + + 0243@xmath67026 & 6820 ( 160 ) & 8.18 ( 0.15 ) & h & 0.70 ( 0.10 ) & 13.90 ( 0.23 ) & @xmath673.61 ( 0.09 ) & 2.27 ( 0.78 ) & + 0245@xmath68541 & 5280 ( 120 ) & 8.28 ( 0.05 ) & h & 0.76 ( 0.03 ) & 15.28 ( 0.07 ) & @xmath674.11 ( 0.03 ) & 6.93 ( 0.30 ) & + 0257@xmath68080 & 6680 ( 150 ) & 7.96 ( 0.16 ) & h & 0.57 ( 0.09 ) & 13.68 ( 0.21 ) & @xmath673.51 ( 0.08 ) & 1.60 ( 0.38 ) & 2 + 0324@xmath68738 & 4650 ( 50 ) & 7.10 ( 0.42 ) & he&0.19 ( 0.10 ) & 14.67 ( 0.33 ) & @xmath673.77 ( 0.14 ) & 1.97 ( 0.74 ) & 6 + 0326@xmath67273 & 7200 ( 200 ) & 7.53 ( 0.40 ) & h & 0.35 ( 0.15 ) & 12.80 ( 0.52 ) & @xmath673.15 ( 0.21 ) & 0.82 ( 0.30 ) & 6 + + 0341@xmath68182 & 6900 ( 170 ) & 8.16 ( 0.09 ) & he&0.67 ( 0.05 ) & 13.80 ( 0.12 ) & @xmath673.58 ( 0.05 ) & 2.24 ( 0.35 ) & 4 + 0357@xmath68081 & 5490 ( 130 ) & 8.02 ( 0.11 ) & h & 0.60 ( 0.06 ) & 14.66 ( 0.14 ) & @xmath673.90 ( 0.06 ) & 3.40 ( 0.87 ) & + 0426@xmath68588 & 7120 ( 180 ) & 8.17 ( 0.01 ) & he&0.68 ( 0.00 ) & 13.71 ( 0.01 ) & @xmath673.54 ( 0.01 ) & 2.08 ( 0.02 ) & + 0433@xmath68270 & 5620 ( 110 ) & 8.14 ( 0.07 ) & h & 0.67 ( 0.05 ) & 14.71 ( 0.10 ) & @xmath673.92 ( 0.04 ) & 4.07 ( 0.69 ) & + 0435@xmath67088 & 6620 ( 160 ) & 8.08 ( 0.04 ) & he&0.62 ( 0.02 ) & 13.86 ( 0.05 ) & @xmath673.61 ( 0.02 ) & 2.22 ( 0.17 ) & 4 + + 0440@xmath68510 & 8710 ( 220 ) & 7.96 ( 0.38 ) & h & 0.57 ( 0.21 ) & 12.62 ( 0.53 ) & @xmath673.05 ( 0.22 ) & 0.82 ( 0.39 ) & + 0503@xmath67174 & 5300 ( 120 ) & 7.60 ( 0.16 ) & h & 0.37 ( 0.07 ) & 14.31 ( 0.19 ) & @xmath673.74 ( 0.07 ) & 1.86 ( 0.33 ) & 2 + 0518@xmath68333 & 7780 ( 180 ) & 7.14 ( 0.24 ) & h & 0.24 ( 0.06 ) & 11.98 ( 0.31 ) & @xmath672.81 ( 0.12 ) & 0.52 ( 0.09 ) & 6 + 0548@xmath67001 & 6400 ( 140 ) & 8.32 ( 0.04 ) & he&0.78 ( 0.03 ) & 14.34 ( 0.07 ) & @xmath673.81 ( 0.02 ) & 3.90 ( 0.22 ) & 2 , 4 + 0551@xmath68468 & 5380 ( 120 ) & 8.01 ( 0.08 ) & h & 0.59 ( 0.04 ) & 14.78 ( 0.10 ) & @xmath673.92 ( 0.04 ) & 3.75 ( 0.73 ) & + 0552@xmath67041 & 5060 ( 60 ) & 8.31 ( 0.02 ) & he&0.77 ( 0.01 ) & 15.42 ( 0.03 ) & @xmath674.21 ( 0.02 ) & 6.92 ( 0.05 ) & 5 + 0553@xmath68053 & 5790 ( 110 ) & 8.20 ( 0.05 ) & h & 0.71 ( 0.03 ) & 14.65 ( 0.06 ) & @xmath673.90 ( 0.02 ) & 4.11 ( 0.35 ) & 2 + 0618@xmath68067 & 5940 ( 120 ) & 8.27 ( 0.14 ) & h & 0.76 ( 0.09 ) & 14.64 ( 0.21 ) & @xmath673.90 ( 0.09 ) & 4.27 ( 0.90 ) & + 0644@xmath68025 & 7410 ( 180 ) & 8.66 ( 0.12 ) & h & 1.01 ( 0.07 ) & 14.38 ( 0.22 ) & @xmath673.78 ( 0.08 ) & 3.79 ( 0.23 ) & + 0648@xmath68641 & 5780 ( 120 ) & 7.75 ( 0.30 ) & h & 0.45 ( 0.14 ) & 14.04 ( 0.38 ) & @xmath673.66 ( 0.15 ) & 1.74 ( 0.65 ) & + + 0651@xmath67479 & 4660 ( 40 ) & 8.02 ( 0.16 ) & he&0.58 ( 0.10 ) & 15.57 ( 0.22 ) & @xmath674.19 ( 0.09 ) & 7.00 ( 1.24 ) & + 0654@xmath68027 & 9450 ( 340 ) & 8.51 ( 0.18 ) & he&0.91 ( 0.12 ) & 13.24 ( 0.31 ) & @xmath673.26 ( 0.12 ) & 1.79 ( 0.56 ) & + 0657@xmath68320 & 4990 ( 130 ) & 8.07 ( 0.03 ) & h & 0.62 ( 0.02 ) & 15.26 ( 0.04 ) & @xmath674.09 ( 0.02 ) & 6.55 ( 0.27 ) & + 0659@xmath67064 & 6520 ( 150 ) & 8.71 ( 0.36 ) & h & 1.04 ( 0.22 ) & 14.97 ( 0.67 ) & @xmath674.04 ( 0.27 ) & 4.81 ( 0.60 ) & + 0706@xmath68377 & 6990 ( 180 ) & 8.14 ( 0.09 ) & he&0.66 ( 0.05 ) & 13.71 ( 0.13 ) & @xmath673.55 ( 0.05 ) & 2.09 ( 0.32 ) & 4 + + 0727@xmath68482a & 5020 ( 120 ) & 7.92 ( 0.02 ) & h & 0.53 ( 0.01 ) & 15.03 ( 0.02 ) & @xmath674.00 ( 0.01 ) & 4.60 ( 0.22 ) & 6 + 0727@xmath68482b & 5000 ( 130 ) & 8.12 ( 0.02 ) & h & 0.66 ( 0.01 ) & 15.33 ( 0.02 ) & @xmath674.12 ( 0.01 ) & 7.02 ( 0.14 ) & 6 + 0738@xmath67172 & 7710 ( 220 ) & 8.09 ( 0.03 ) & he&0.63 ( 0.02 ) & 13.31 ( 0.05 ) & @xmath673.35 ( 0.02 ) & 1.45 ( 0.05 ) & + 0743@xmath67340 & 4740 ( 50 ) & 7.97 ( 0.09 ) & he&0.55 ( 0.05 ) & 15.40 ( 0.12 ) & @xmath674.14 ( 0.04 ) & 6.36 ( 0.96 ) & + 0747@xmath68073a & 4850 ( 50 ) & 8.04 ( 0.02 ) & he&0.59 ( 0.01 ) & 15.32 ( 0.03 ) & @xmath674.13 ( 0.01 ) & 6.71 ( 0.15 ) & + + 0747@xmath68073b & 4210 ( 90 ) & 7.69 ( 0.03 ) & h & 0.40 ( 0.01 ) & 15.65 ( 0.03 ) & @xmath674.19 ( 0.01 ) & 5.27 ( 0.23 ) & + 0751@xmath68578 & 9770 ( 360 ) & 8.01 ( 0.21 ) & he&0.58 ( 0.12 ) & 12.35 ( 0.29 ) & @xmath672.89 ( 0.12 ) & 0.71 ( 0.20 ) & + 0752@xmath67676 & 5730 ( 110 ) & 8.21 ( 0.09 ) & h & 0.72 ( 0.06 ) & 14.70 ( 0.13 ) & @xmath673.93 ( 0.06 ) & 4.36 ( 0.70 ) & + 0752@xmath68365 & 7700 ( 190 ) & 8.19 ( 0.21 ) & h & 0.71 ( 0.13 ) & 13.44 ( 0.32 ) & @xmath673.40 ( 0.13 ) & 1.60 ( 0.69 ) & + 0802@xmath68386&11070 ( 480 ) & 8.31 ( 0.18 ) & he&0.78 ( 0.11 ) & 12.46 ( 0.27 ) & @xmath672.85 ( 0.11 ) & 0.76 ( 0.18 ) & 5 + + 0816@xmath68387 & 7570 ( 190 ) & 8.19 ( 0.14 ) & h & 0.71 ( 0.09 ) & 13.52 ( 0.22 ) & @xmath673.43 ( 0.09 ) & 1.67 ( 0.51 ) & + 0827@xmath68328 & 7270 ( 180 ) & 8.39 ( 0.11 ) & h & 0.85 ( 0.07 ) & 13.99 ( 0.18 ) & @xmath673.63 ( 0.08 ) & 3.06 ( 0.58 ) & + 0839@xmath67327 & 8930 ( 230 ) & 7.70 ( 0.14 ) & h & 0.44 ( 0.07 ) & 12.16 ( 0.19 ) & @xmath672.86 ( 0.07 ) & 0.57 ( 0.08 ) & 6 + 0856@xmath68331&10390 ( 410 ) & 8.84 ( 0.07 ) & he&1.11 ( 0.04 ) & 13.60 ( 0.15 ) & @xmath673.33 ( 0.06 ) & 2.07 ( 0.08 ) & 4 + 0912@xmath68536 & 7160 ( 190 ) & 8.28 ( 0.03 ) & he&0.75 ( 0.02 ) & 13.78 ( 0.04 ) & @xmath673.59 ( 0.02 ) & 2.54 ( 0.16 ) & 2 + + 0913@xmath68442 & 8490 ( 220 ) & 8.19 ( 0.17 ) & h & 0.71 ( 0.11 ) & 13.06 ( 0.25 ) & @xmath673.23 ( 0.10 ) & 1.22 ( 0.34 ) & + 0930@xmath68294 & 8330 ( 220 ) & 8.38 ( 0.20 ) & h & 0.84 ( 0.13 ) & 13.45 ( 0.32 ) & @xmath673.38 ( 0.12 ) & 1.94 ( 0.74 ) & + 0946@xmath68534 & 8760 ( 290 ) & 8.45 ( 0.11 ) & he&0.87 ( 0.07 ) & 13.37 ( 0.18 ) & @xmath673.35 ( 0.07 ) & 1.97 ( 0.41 ) & 4 + 0955@xmath68247 & 8670 ( 220 ) & 8.27 ( 0.15 ) & h & 0.77 ( 0.10 ) & 13.12 ( 0.24 ) & @xmath673.24 ( 0.09 ) & 1.31 ( 0.40 ) & + 1012@xmath68083 & 6750 ( 150 ) & 8.02 ( 0.18 ) & h & 0.60 ( 0.10 ) & 13.72 ( 0.25 ) & @xmath673.53 ( 0.09 ) & 1.72 ( 0.52 ) & + + 1019@xmath68637 & 6780 ( 160 ) & 7.98 ( 0.09 ) & h & 0.58 ( 0.05 ) & 13.63 ( 0.13 ) & @xmath673.50 ( 0.05 ) & 1.59 ( 0.22 ) & + 1039@xmath68145 & 7280 ( 200 ) & 7.93 ( 0.26 ) & he&0.53 ( 0.14 ) & 13.24 ( 0.35 ) & @xmath673.36 ( 0.14 ) & 1.32 ( 0.47 ) & 1 + 1043@xmath67188 & 6190 ( 200 ) & 8.09 ( 0.17 ) & he&0.63 ( 0.11 ) & 14.30 ( 0.25 ) & @xmath673.74 ( 0.10 ) & 2.85 ( 1.03 ) & 3 + 1055@xmath67072 & 7420 ( 200 ) & 8.42 ( 0.06 ) & he&0.85 ( 0.04 ) & 13.91 ( 0.09 ) & @xmath673.62 ( 0.04 ) & 3.01 ( 0.22 ) & + 1108@xmath68207 & 4650 ( 160 ) & 8.07 ( 0.11 ) & h & 0.62 ( 0.07 ) & 15.62 ( 0.15 ) & @xmath674.21 ( 0.06 ) & 7.71 ( 0.98 ) & + 1115@xmath67029&10120 ( 490 ) & 8.13 ( 0.30 ) & he&0.66 ( 0.18 ) & 12.44 ( 0.43 ) & @xmath672.90 ( 0.17 ) & 0.76 ( 0.31 ) & 4 + 1121@xmath68216 & 7490 ( 180 ) & 8.20 ( 0.05 ) & h & 0.72 ( 0.03 ) & 13.57 ( 0.08 ) & @xmath673.45 ( 0.03 ) & 1.76 ( 0.22 ) & + 1124@xmath67296 & 9440 ( 250 ) & 7.10 ( 0.15 ) & h & 0.23 ( 0.03 ) & 11.09 ( 0.23 ) & @xmath672.43 ( 0.10 ) & 0.30 ( 0.04 ) & 6 + 1136@xmath67286 & 4490 ( 80 ) & 9.02 ( 0.05 ) & he&1.19 ( 0.02 ) & 17.47 ( 0.09 ) & @xmath674.94 ( 0.03 ) & 4.96 ( 0.27 ) & + 1142@xmath67645 & 8490 ( 270 ) & 8.27 ( 0.05 ) & he&0.75 ( 0.03 ) & 13.19 ( 0.07 ) & @xmath673.28 ( 0.02 ) & 1.44 ( 0.13 ) & 4 + + 1147@xmath68255 & 9790 ( 270 ) & 7.92 ( 0.23 ) & h & 0.55 ( 0.13 ) & 12.11 ( 0.33 ) & @xmath672.82 ( 0.13 ) & 0.57 ( 0.17 ) & + 1154@xmath68186 & 7630 ( 220 ) & 8.06 ( 0.16 ) & he&0.61 ( 0.10 ) & 13.23 ( 0.23 ) & @xmath673.35 ( 0.09 ) & 1.44 ( 0.32 ) & 1 + 1208@xmath68576 & 5880 ( 120 ) & 7.95 ( 0.15 ) & h & 0.56 ( 0.09 ) & 14.23 ( 0.20 ) & @xmath673.73 ( 0.08 ) & 2.18 ( 0.60 ) & + 1215@xmath68323 & 7100 ( 210 ) & 8.68 ( 0.13 ) & he&1.02 ( 0.08 ) & 14.54 ( 0.24 ) & @xmath673.88 ( 0.10 ) & 3.91 ( 0.10 ) & 1 + 1236@xmath67495&11550 ( 470 ) & 8.63 ( 0.18 ) & h & 1.00 ( 0.11 ) & 12.73 ( 0.34 ) & @xmath672.98 ( 0.13 ) & 1.24 ( 0.44 ) & 7 + + 1244@xmath68149&10280 ( 310 ) & 7.88 ( 0.42 ) & h & 0.53 ( 0.21 ) & 11.87 ( 0.60 ) & @xmath672.71 ( 0.24 ) & 0.48 ( 0.24 ) & + 1247@xmath68550 & 4050 ( 70 ) & 7.57 ( 0.03 ) & h & 0.35 ( 0.01 ) & 15.78 ( 0.04 ) & @xmath674.20 ( 0.01 ) & 4.63 ( 0.31 ) & + 1257@xmath68037 & 5590 ( 110 ) & 8.16 ( 0.09 ) & h & 0.68 ( 0.06 ) & 14.74 ( 0.14 ) & @xmath673.94 ( 0.05 ) & 4.32 ( 0.88 ) & + 1257@xmath68278 & 8540 ( 240 ) & 7.97 ( 0.22 ) & h & 0.58 ( 0.13 ) & 12.71 ( 0.31 ) & @xmath673.09 ( 0.12 ) & 0.87 ( 0.26 ) & + 1300@xmath68263 & 4320 ( 100 ) & 8.07 ( 0.18 ) & h & 0.62 ( 0.12 ) & 16.04 ( 0.25 ) & @xmath674.34 ( 0.11 ) & 8.73 ( 1.47 ) & + + 1310@xmath67472 & 4220 ( 80 ) & 8.12 ( 0.05 ) & h & 0.65 ( 0.04 ) & 16.24 ( 0.08 ) & @xmath674.41 ( 0.03 ) & 9.34 ( 0.34 ) & + 1313@xmath67198 & 5330 ( 100 ) & 8.33 ( 0.06 ) & he&0.79 ( 0.04 ) & 15.18 ( 0.10 ) & @xmath674.14 ( 0.04 ) & 6.47 ( 0.19 ) & + 1325@xmath68581 & 6810 ( 160 ) & 8.14 ( 0.31 ) & h & 0.68 ( 0.19 ) & 13.86 ( 0.45 ) & @xmath673.58 ( 0.17 ) & 2.08 ( 1.21 ) & + 1328@xmath68307 & 7320 ( 230 ) & 8.21 ( 0.24 ) & he&0.71 ( 0.15 ) & 13.65 ( 0.35 ) & @xmath673.51 ( 0.14 ) & 2.02 ( 0.87 ) & 5 + 1334@xmath68039 & 5030 ( 120 ) & 7.95 ( 0.05 ) & h & 0.55 ( 0.03 ) & 15.05 ( 0.06 ) & @xmath674.01 ( 0.02 ) & 4.90 ( 0.57 ) & + + 1344@xmath68106 & 7110 ( 170 ) & 8.10 ( 0.11 ) & h & 0.65 ( 0.07 ) & 13.61 ( 0.16 ) & @xmath673.48 ( 0.06 ) & 1.67 ( 0.31 ) & + 1345@xmath68238 & 4590 ( 150 ) & 7.76 ( 0.05 ) & h & 0.44 ( 0.02 ) & 15.30 ( 0.06 ) & @xmath674.07 ( 0.02 ) & 4.57 ( 0.51 ) & + 1418@xmath67088 & 7810 ( 190 ) & 7.43 ( 0.26 ) & h & 0.32 ( 0.09 ) & 12.34 ( 0.34 ) & @xmath672.96 ( 0.13 ) & 0.63 ( 0.12 ) & 6 + 1444@xmath67174 & 4960 ( 60 ) & 8.37 ( 0.08 ) & he&0.81 ( 0.05 ) & 15.63 ( 0.13 ) & @xmath674.29 ( 0.05 ) & 7.19 ( 0.07 ) & + 1455@xmath68298 & 7310 ( 170 ) & 7.66 ( 0.23 ) & h & 0.41 ( 0.11 ) & 12.90 ( 0.31 ) & @xmath673.19 ( 0.12 ) & 0.89 ( 0.21 ) & 6 + + 1503@xmath67070 & 6990 ( 160 ) & 8.17 ( 0.21 ) & h & 0.70 ( 0.13 ) & 13.81 ( 0.31 ) & @xmath673.56 ( 0.12 ) & 2.02 ( 0.91 ) & 2 + 1606@xmath68422&11320 ( 570 ) & 7.12 ( 0.23 ) & h & 0.25 ( 0.05 ) & 10.55 ( 0.35 ) & @xmath672.11 ( 0.13 ) & 0.18 ( 0.04 ) & 6 + 1609@xmath68135 & 9080 ( 250 ) & 8.75 ( 0.10 ) & h & 1.07 ( 0.06 ) & 13.79 ( 0.19 ) & @xmath673.50 ( 0.07 ) & 2.71 ( 0.17 ) & + 1625@xmath68093 & 6870 ( 170 ) & 8.44 ( 0.12 ) & h & 0.88 ( 0.07 ) & 14.30 ( 0.19 ) & @xmath673.76 ( 0.07 ) & 3.74 ( 0.48 ) & + 1626@xmath68368 & 8640 ( 280 ) & 8.03 ( 0.05 ) & he&0.60 ( 0.03 ) & 12.82 ( 0.07 ) & @xmath673.12 ( 0.03 ) & 1.02 ( 0.07 ) & + + 1633@xmath68433 & 6650 ( 150 ) & 8.14 ( 0.07 ) & h & 0.68 ( 0.04 ) & 13.94 ( 0.10 ) & @xmath673.63 ( 0.04 ) & 2.28 ( 0.34 ) & + 1633@xmath68572 & 6180 ( 240 ) & 8.09 ( 0.06 ) & he&0.63 ( 0.03 ) & 14.19 ( 0.08 ) & @xmath673.74 ( 0.03 ) & 2.84 ( 0.37 ) & 3 + 1635@xmath68137 & 6990 ( 190 ) & 8.29 ( 0.26 ) & h & 0.78 ( 0.16 ) & 13.97 ( 0.40 ) & @xmath673.63 ( 0.15 ) & 2.77 ( 1.15 ) & + 1637@xmath68335 & 9940 ( 280 ) & 8.13 ( 0.13 ) & h & 0.68 ( 0.08 ) & 12.37 ( 0.20 ) & @xmath672.91 ( 0.08 ) & 0.74 ( 0.14 ) & + 1639@xmath68537 & 7510 ( 210 ) & 8.14 ( 0.11 ) & he&0.67 ( 0.07 ) & 13.43 ( 0.16 ) & @xmath673.42 ( 0.07 ) & 1.64 ( 0.27 ) & 2 + 1655@xmath68215 & 9180 ( 230 ) & 7.87 ( 0.12 ) & h & 0.53 ( 0.06 ) & 12.30 ( 0.16 ) & @xmath672.91 ( 0.07 ) & 0.64 ( 0.09 ) & + 1656@xmath67062 & 5520 ( 100 ) & 8.03 ( 0.04 ) & h & 0.60 ( 0.03 ) & 14.58 ( 0.06 ) & @xmath673.89 ( 0.03 ) & 3.34 ( 0.38 ) & + 1705@xmath68030 & 7050 ( 180 ) & 8.35 ( 0.13 ) & he&0.80 ( 0.09 ) & 13.98 ( 0.21 ) & @xmath673.66 ( 0.08 ) & 3.11 ( 0.62 ) & 5 + 1716@xmath68020&13470 ( 900 ) & 8.11 ( 0.13 ) & h & 0.67 ( 0.08 ) & 11.60 ( 0.20 ) & @xmath672.37 ( 0.09 ) & 0.32 ( 0.06 ) & + 1733@xmath67544 & 6520 ( 150 ) & 8.13 ( 0.03 ) & h & 0.67 ( 0.02 ) & 13.99 ( 0.04 ) & @xmath673.65 ( 0.01 ) & 2.37 ( 0.14 ) & + + 1736@xmath68052 & 8890 ( 230 ) & 8.11 ( 0.25 ) & h & 0.67 ( 0.15 ) & 12.77 ( 0.37 ) & @xmath673.10 ( 0.14 ) & 0.97 ( 0.35 ) & + 1748@xmath68708 & 5590 ( 90 ) & 8.36 ( 0.02 ) & he&0.81 ( 0.01 ) & 15.21 ( 0.03 ) & @xmath674.07 ( 0.02 ) & 5.86 ( 0.06 ) & 2 , 3 ? + 1756@xmath68827 & 7270 ( 330 ) & 7.98 ( 0.07 ) & h & 0.58 ( 0.04 ) & 13.37 ( 0.10 ) & @xmath673.38 ( 0.04 ) & 1.35 ( 0.13 ) & + 1811@xmath68327a & 7380 ( 180 ) & 7.65 ( 0.37 ) & h & 0.41 ( 0.16 ) & 12.85 ( 0.49 ) & @xmath673.17 ( 0.19 ) & 0.86 ( 0.33 ) & + 1811@xmath68327b & 6480 ( 160 ) & 7.75 ( 0.37 ) & h & 0.45 ( 0.18 ) & 13.52 ( 0.49 ) & @xmath673.45 ( 0.20 ) & 1.32 ( 0.60 ) & + + 1818@xmath68126 & 6340 ( 130 ) & 7.28 ( 0.41 ) & h & 0.26 ( 0.12 ) & 13.04 ( 0.50 ) & @xmath673.26 ( 0.20 ) & 0.93 ( 0.26 ) & 6 + 1820@xmath68609 & 4780 ( 140 ) & 7.83 ( 0.09 ) & h & 0.48 ( 0.05 ) & 15.16 ( 0.11 ) & @xmath674.03 ( 0.04 ) & 4.68 ( 0.95 ) & + 1824@xmath68040&12240 ( 680 ) & 7.00 ( 0.16 ) & h & 0.23 ( 0.03 ) & 10.19 ( 0.28 ) & @xmath671.89 ( 0.11 ) & 0.12 ( 0.03 ) & 6 + 1826@xmath67045 & 9480 ( 350 ) & 7.94 ( 0.17 ) & h & 0.57 ( 0.09 ) & 12.29 ( 0.24 ) & @xmath672.89 ( 0.10 ) & 0.64 ( 0.14 ) & + 1829@xmath68547 & 6280 ( 140 ) & 8.50 ( 0.11 ) & h & 0.90 ( 0.07 ) & 14.69 ( 0.18 ) & @xmath673.96 ( 0.08 ) & 4.76 ( 0.24 ) & 2 , 8 + + 1831@xmath68197 & 7590 ( 210 ) & 7.61 ( 0.48 ) & he&0.37 ( 0.19 ) & 12.67 ( 0.60 ) & @xmath673.12 ( 0.24 ) & 0.80 ( 0.40 ) & 4 + 1840@xmath68042 & 9090 ( 340 ) & 8.19 ( 0.12 ) & h & 0.71 ( 0.08 ) & 12.81 ( 0.18 ) & @xmath673.11 ( 0.08 ) & 1.02 ( 0.18 ) & + 1855@xmath68338&11240 ( 430 ) & 8.19 ( 0.20 ) & h & 0.72 ( 0.13 ) & 12.06 ( 0.32 ) & @xmath672.73 ( 0.12 ) & 0.58 ( 0.18 ) & 7 + 1900@xmath68705&12070 ( 990 ) & 8.58 ( 0.03 ) & he&0.95 ( 0.02 ) & 12.68 ( 0.06 ) & @xmath672.88 ( 0.03 ) & 0.94 ( 0.09 ) & 2 + 1917@xmath68386 & 6390 ( 140 ) & 8.28 ( 0.05 ) & he&0.75 ( 0.04 ) & 14.27 ( 0.09 ) & @xmath673.79 ( 0.03 ) & 3.68 ( 0.32 ) & 3 ? + + 1953@xmath67011 & 7920 ( 200 ) & 8.23 ( 0.05 ) & h & 0.74 ( 0.03 ) & 13.41 ( 0.07 ) & @xmath673.38 ( 0.03 ) & 1.63 ( 0.16 ) & 2 + 2002@xmath67110 & 4800 ( 50 ) & 8.31 ( 0.02 ) & he&0.77 ( 0.01 ) & 15.76 ( 0.03 ) & @xmath674.31 ( 0.01 ) & 7.35 ( 0.04 ) & + 2011@xmath68065 & 6400 ( 150 ) & 8.13 ( 0.06 ) & he&0.65 ( 0.04 ) & 14.03 ( 0.09 ) & @xmath673.69 ( 0.03 ) & 2.73 ( 0.38 ) & 1 + 2048@xmath68263 & 5200 ( 110 ) & 7.31 ( 0.12 ) & h & 0.26 ( 0.04 ) & 14.12 ( 0.15 ) & @xmath673.64 ( 0.06 ) & 1.53 ( 0.15 ) & 6 + 2054@xmath67050 & 4620 ( 40 ) & 8.09 ( 0.12 ) & he&0.62 ( 0.08 ) & 15.74 ( 0.17 ) & @xmath674.24 ( 0.07 ) & 7.49 ( 0.56 ) & + + 2059@xmath68190 & 6840 ( 160 ) & 7.86 ( 0.28 ) & h & 0.51 ( 0.15 ) & 13.44 ( 0.37 ) & @xmath673.42 ( 0.15 ) & 1.33 ( 0.48 ) & + 2059@xmath68247 & 6070 ( 130 ) & 8.17 ( 0.17 ) & h & 0.69 ( 0.11 ) & 14.36 ( 0.25 ) & @xmath673.80 ( 0.10 ) & 3.32 ( 1.12 ) & + 2059@xmath68316&10080 ( 390 ) & 8.08 ( 0.18 ) & he&0.63 ( 0.11 ) & 12.35 ( 0.26 ) & @xmath672.87 ( 0.11 ) & 0.72 ( 0.17 ) & 4 + 2105@xmath67820&10200 ( 290 ) & 8.23 ( 0.21 ) & h & 0.75 ( 0.13 ) & 12.45 ( 0.33 ) & @xmath672.93 ( 0.13 ) & 0.81 ( 0.27 ) & + 2107@xmath67216 & 5830 ( 140 ) & 8.40 ( 0.05 ) & h & 0.85 ( 0.03 ) & 14.93 ( 0.08 ) & @xmath674.02 ( 0.03 ) & 5.24 ( 0.23 ) & + + 2111@xmath68261 & 8120 ( 190 ) & 7.42 ( 0.19 ) & h & 0.32 ( 0.07 ) & 12.16 ( 0.26 ) & @xmath672.88 ( 0.11 ) & 0.57 ( 0.08 ) & 6 + 2136@xmath68229 & 9480 ( 250 ) & 7.85 ( 0.20 ) & h & 0.52 ( 0.11 ) & 12.13 ( 0.28 ) & @xmath672.84 ( 0.12 ) & 0.58 ( 0.14 ) & + 2140@xmath68207 & 8860 ( 300 ) & 8.07 ( 0.06 ) & he&0.62 ( 0.04 ) & 12.75 ( 0.09 ) & @xmath673.09 ( 0.04 ) & 1.00 ( 0.07 ) & 4 + 2207@xmath68142 & 7620 ( 190 ) & 8.24 ( 0.16 ) & h & 0.75 ( 0.10 ) & 13.57 ( 0.24 ) & @xmath673.45 ( 0.10 ) & 1.86 ( 0.64 ) & + 2246@xmath68223&10330 ( 300 ) & 8.57 ( 0.09 ) & h & 0.97 ( 0.06 ) & 12.99 ( 0.17 ) & @xmath673.14 ( 0.07 ) & 1.56 ( 0.33 ) & + 2248@xmath68293 & 5580 ( 110 ) & 7.53 ( 0.15 ) & h & 0.35 ( 0.06 ) & 13.94 ( 0.19 ) & @xmath673.61 ( 0.08 ) & 1.50 ( 0.21 ) & 6 + 2251@xmath67070 & 4580 ( 60 ) & 8.38 ( 0.05 ) & he&0.82 ( 0.03 ) & 16.17 ( 0.08 ) & @xmath674.43 ( 0.03 ) & 7.75 ( 0.01 ) & 5 + 2253@xmath67081 & 6770 ( 180 ) & 8.41 ( 0.19 ) & h & 0.86 ( 0.13 ) & 14.30 ( 0.32 ) & @xmath673.76 ( 0.12 ) & 3.75 ( 0.94 ) & + 2311@xmath67068 & 7980 ( 240 ) & 8.27 ( 0.17 ) & he&0.75 ( 0.11 ) & 13.40 ( 0.26 ) & @xmath673.39 ( 0.11 ) & 1.74 ( 0.58 ) & 4 + 2312@xmath67024 & 6840 ( 160 ) & 8.41 ( 0.22 ) & he&0.84 ( 0.14 ) & 14.18 ( 0.34 ) & @xmath673.75 ( 0.14 ) & 3.62 ( 0.85 ) & 5 + + 2316@xmath67064 & 4720 ( 50 ) & 8.17 ( 0.17 ) & he&0.68 ( 0.11 ) & 15.69 ( 0.25 ) & @xmath674.25 ( 0.10 ) & 7.30 ( 0.52 ) & + 2329@xmath68267 & 9400 ( 240 ) & 8.02 ( 0.28 ) & h & 0.61 ( 0.16 ) & 12.40 ( 0.40 ) & @xmath672.95 ( 0.15 ) & 0.73 ( 0.27 ) & + 2345@xmath67447 & 5400 ( 80 ) & 8.72 ( 0.07 ) & he&1.04 ( 0.04 ) & 15.77 ( 0.13 ) & @xmath674.38 ( 0.05 ) & 6.07 ( 0.24 ) & + 2347@xmath68292 & 5810 ( 120 ) & 7.82 ( 0.15 ) & h & 0.48 ( 0.08 ) & 14.10 ( 0.19 ) & @xmath673.69 ( 0.08 ) & 1.89 ( 0.37 ) & + 2352@xmath68401 & 8260 ( 260 ) & 7.99 ( 0.23 ) & he&0.57 ( 0.13 ) & 12.88 ( 0.32 ) & @xmath673.17 ( 0.12 ) & 1.07 ( 0.32 ) & 4 + + | a detailed photometric and spectroscopic analysis of cool ( @xmath0 k ) white dwarf stars is presented .
the sample has been drawn from the yale parallax catalog and from a proper - motion survey in the southern hemisphere .
optical @xmath1 and infrared @xmath2 photometry , as well as spectroscopy at h@xmath3 , have been secured for a sample of 152 white dwarfs .
the discovery of seven new da white dwarfs , two new dq white dwarfs , one new magnetic white dwarf , and three weak magnetic white dwarf candidates , is reported .
our sample also identifies 19 known or suspected double degenerates .
the photometric energy distributions , the h@xmath3 line profiles , and the trigonometric parallax measurements are combined and compared against the predictions of model atmosphere calculations to determine the effective temperature and the radius of each object in the sample , and also to constrain the atmospheric composition .
new evolutionary sequences with carbon / oxygen cores with thin and thick hydrogen layers are used to derive stellar masses and ages .
the results are used to improve our understanding of the chemical evolution of cool white dwarfs .
we confirm the existence of a range in effective temperature between @xmath45000 and 6000 k where almost all white dwarfs have hydrogen - rich atmospheres .
our sample shows little evidence for mixed h / he white dwarfs , with the exception of two helium - rich da stars , and four ( possibly five ) c@xmath5h white dwarfs which have been interpreted as having mixed h / he / c atmospheres .
the observed sequence of dq stars is found to terminate abruptly near 6500 k , below which they are believed to turn into c@xmath5h stars .
true dc stars slightly above this temperature are found to exhibit hydrogen - like energy distributions despite the lack of h@xmath3 absorption features .
the mean mass of our complete sample is 0.65 @xmath6 with a dispersion of @xmath7 @xmath6 .
attempts to interpret the chemical evolution of cool white dwarfs show the problem to be complex .
convective mixing is called upon to account for the increase of the non - da to da ratio below 12,000 k , as well as the reappearance of helium - rich stars below @xmath8 k. the possible presence of helium in cool da stars , the existence of the non - da gap , and the nature of the peculiar dc stars are also explained in terms of convective mixing , although our understanding of how this mechanism works needs to be revised in order to account for these observations .
given this chemical evolution uncertainty , it is not clear whether thick or thin hydrogen layer models should be used to determine cooling ages .
the oldest object in our sample is @xmath47.9 gyr or @xmath49.7 gyr old depending on whether thin or thick hydrogen layer models are used , respectively . |
field - theoretic tools and the _ renormalisation group _ ( rg ) method have had a tremendous impact in our understanding of the _ universal power laws _ that emerge near equilibrium critical points ( see , e.g. , refs .
@xcite ) , including the associated _ dynamic critical phenomena
our goal here is to similarly describe the scaling properties of systems driven far from thermal equilibrium , which either undergo a _ continuous nonequilibrium phase transition _ or display _ generic scale invariance_. we are then confronted with capturing the ( stochastic ) dynamics of the long - wavelength modes of the ` slow ' degrees of freedom , namely the order parameter for the transition , any conserved quantities , and perhaps additional relevant variables . in these lecture notes ,
i aim to briefly describe how a representation in terms of a _ field theory action _ can be obtained for ( 1 ) general nonlinear langevin stochastic differential equations @xcite ; and ( 2 ) for master equations governing classical particle reaction diffusion systems @xcite .
i will then demonstrate how the dynamic ( perturbative ) rg can be employed to derive the asymptotic _ scaling laws _ in stochastic dynamical systems ; to infer the _ upper critical dimension _
@xmath0 ( for dimensions @xmath1 , fluctuations strongly affect the universal scaling properties ) ; and to systematically compute the _ critical exponents _ as well as to determine further universal properties in various intriguing dynamical model systems both near and far from equilibrium .
( for considerably more details , especially on the more technical aspects , the reader is referred to ref .
@xcite . )
the vicinity of a _
critical point _ is characterised by _
strong correlations _ and _ large fluctuations_. the system under investigation is then behaving in a highly cooperative manner , and as a consequence , the standard approximative methods of statistical mechanics , namely perturbation or cluster expansions that assume either weak interactions or short - range correlations , fail . upon approaching an equilibrium continuous ( second - order ) phase transition , i.e. , for @xmath2 , where @xmath3 measures the deviation from the critical temperature @xmath4 , the thermal fluctuations of the _ order parameter _ @xmath5 ( which characterises the different thermodynamic phases , usually chosen such that the thermal average @xmath6 vanishes in the high - temperature ` disordered ' phase ) are , in the thermodynamic limit , governed by a diverging length scale @xmath7 here , we have defined the _ correlation length _ via the typically exponential decay of the static _ cumulant _ or _ connected two - point correlation function _
@xmath8 , and @xmath9 denotes the correlation length _
critical exponent_. as @xmath10 , @xmath11 , which entails the absence of any characteristic length scale for the order parameter fluctuations at criticality .
hence we expect the _ critical correlations _ to follow a power law @xmath12 in @xmath13 dimensions , which defines the _ fisher exponent _ @xmath14 .
the following _ scaling ansatz _ generalises this power law to @xmath15 , but still in the vicinity of the critical point , @xmath16 with two distinct regular _ scaling functions _
@xmath17 for @xmath18 and @xmath19 for @xmath20 , respectively . for
its fourier transform @xmath21 , one obtains the corresponding scaling form @xmath22 with new scaling functions @xmath23 . as we will see in section [ ssec : i5 ]
, there are only _ two _ independent static critical exponents .
consequently , it must be possible to use the static scaling hypothesis ( [ i1cumx ] ) or ( [ i1cumq ] ) , along with the definition ( [ i1nuex ] ) , to express the exponents describing the thermodynamic singularities near a second - order phase transition in terms of @xmath9 and @xmath14 through _ scaling laws_. for example , the order parameter in the low - temperature phase ( @xmath24 ) is expected to grow as @xmath25 .
let us consider eq .
( [ i1cumx ] ) in the limit @xmath26 . in order for the @xmath27 dependence to cancel , @xmath28 for large @xmath29 , and therefore @xmath30 . on the other hand , @xmath31 for @xmath20
; thus we identify the order parameter critical exponent through the _ hyperscaling relation _
@xmath32 let us next consider the isothermal static _
@xmath33 , which according to the equilibrium fluctuation
response theorem is given in terms of the spatial integral of the correlation function @xmath34 : @xmath35 .
but @xmath36 as @xmath37 to ensure nonsingular behaviour , whence @xmath38 , and upon defining the associated thermodynamic critical exponent @xmath39 via @xmath40 , we obtain the scaling relation @xmath41 the scaling laws ( [ i1cumx ] ) , ( [ i1cumq ] ) as well as scaling relations such as ( [ i1opex ] ) and ( [ i1scex ] ) can be put on solid foundations by means of the rg procedure , based on an _ effective _ long - wavelength hamiltonian @xmath42 $ ] , a functional of @xmath43 , that captures the essential physics of the problem , namely the relevant symmetries in order parameter and real space , and the existence of a continuous phase transition . the probability of finding a configuration @xmath43 at given temperature @xmath44 is then given by the canonical distribution @xmath45
\propto \exp\left ( -{\cal h}[s ] / k_{\rm b}t \right ) \ . \label{i1eqpd}\ ] ] for example , the mathematical description of the critical phenomena for an @xmath46-symmetric order parameter field @xmath47 , with vector index @xmath48 , is based on the _ landau ginzburg wilson functional _
@xmath49 & = & \int \ ! \d^dx
\sum_\alpha \biggl [ \frac{r}{2 } \ , [ s^\alpha(\vec{x})]^2 + \frac12 \ , [ \nabla s^\alpha(\vec{x})]^2 \nonumber \\ & & \qquad\qquad + \frac{u}{4 ! } \ , [ s^\alpha(\vec{x})]^2 \sum_\beta [ s^\beta(\vec{x})]^2 - h^\alpha(\vec{x } ) \ , s^\alpha(\vec{x } ) \biggr ] \ , \label{i1glwh } \end{aligned}\ ] ] where @xmath50 is the external field thermodynamically conjugate to @xmath47 , @xmath51 denotes the strength of the nonlinearity that drives the phase transformation , and @xmath52 is the control parameter for the transition , i.e. , @xmath53 , where @xmath54 is the ( mean - field ) critical temperature .
spatial variations of the order parameter are energetically suppressed by the term @xmath55 ^ 2 $ ] , and the corresponding positive coefficient has been absorbed into the fields @xmath56 .
we shall , however , not pursue the static theory further here , but instead proceed to a full _ dynamical description _ in terms of nonlinear langevin equations @xcite .
we will then formulate the rg within this dynamic framework , and therein demonstrate the emergence of scaling laws and the computation of critical exponents in a systematic perturbative expansion with respect to the deviation @xmath57 from the upper critical dimension . in order to construct the desired effective stochastic dynamics near a critical point
, we recall that correlated region of size @xmath58 become quite large in the vicinity of the transition .
since the associated relaxation times for such clusters should grow with their extent , one would expect the characteristic time scale for the relaxation of the order parameter fluctuations to increase as well as @xmath10 , namely @xmath59 which introduces the _ dynamic critical exponent _
@xmath60 that encodes the _ critical slowing down _ at the phase transition ; usually @xmath61 . since the typical relaxation rates therefore scale as @xmath62 , we may utilise the static scaling variable @xmath63 to generalise the crucial observation ( [ i1dyex ] ) and formulate a _ dynamic scaling hypothesis _ for the wavevector - dependent dispersion relation of the order parameter fluctuations @xcite , @xmath64 we can then proceed to write down dynamical scaling laws by simply postulating the additional scaling variables @xmath65 or @xmath66 . for example , as an immediate consequence we find for the time - dependent mean order parameter @xmath67 with @xmath68 , but @xmath69 as @xmath70 in order for the @xmath71 dependence to disappear . at the critical point ( @xmath72 ) , this yields the power - law decay @xmath73 , with @xmath74 similarly , the scaling law for the _ dynamic order parameter susceptibility ( response function ) _ becomes @xmath75 which constitutes the dynamical generalisation of eq .
( [ i1cumq ] ) , for @xmath76 . upon applying the _ fluctuation
dissipation theorem _ , valid in thermal equilibrium , we therefrom obtain the _ dynamic correlation function _ @xmath77 and for its fourier transform in real space and time , @xmath78 which reduces to the static limit ( [ i1cumx ] ) if we set @xmath79 .
the critical slowing down of the order parameter fluctuations near the critical point provides us with a natural _ separation of time scales_. assuming ( for now ) that there are no other conserved variables in the system , which would constitute additional slow modes , we may thus resort to a _ coarse - grained _ long - wavelength and long - time description , focusing merely on the order parameter kinetics , while subsuming all other ` fast ' degrees of freedom in _
random ` noise ' _ terms .
this leads us to a _ mesoscopic langevin equation _ for the _ slow _
variables @xmath80 of the form @xmath81(\vec{x},t ) + \zeta^\alpha(\vec{x},t ) \ . \label{i1land}\ ] ] in the simplest case , the _
systematic _ force terms here just represent purely _ relaxational _ dynamics towards the equilibrium configuration @xcite , @xmath82(\vec{x},t ) = - d \
, \frac{\delta { \cal h}[s]}{\delta s^\alpha(x , t ) } \ , \label{i1reld}\ ] ] where @xmath83 represents the relaxation coefficient , and @xmath42 $ ] is again the effective hamiltonian that governs the phase transition , e.g. given by eq .
( [ i1glwh ] ) . for the _ stochastic forces
_ we may assume the most convenient form , and take them to simply represent gaussian white noise with zero mean , @xmath84 , but with their second moment in thermal equilibrium fixed by _
einstein s relation _
@xmath85 as can be verified by means of the associated fokker
planck equation for the time - dependent probability distribution @xmath86 $ ] , eq .
( [ i1eins ] ) guarantees that eventually @xmath87 \to { \cal p}_{\rm eq}[s]$ ] , the canonical distribution ( [ i1eqpd ] ) . the stochastic differential equation ( [ i1land ] ) , with ( [ i1reld ] ) , the hamiltonian ( [ i1glwh ] ) , and the noise correlator ( [ i1eins ] ) , define the _ relaxational model a _ ( according to the classification in ref .
@xcite ) for a nonconserved @xmath46-symmetric order parameter .
if , however , the order parameter is _ conserved _ , we have to consider the associated continuity equation @xmath88 , where typically the conserved current is given by a gradient of the field @xmath56 : @xmath89 ; as a consequence , the order parameter fluctuations will relax _ diffusively _ with diffusion coefficient @xmath83 .
the ensuing _ model b _
@xcite for the relaxational critical dynamics of a conserved order parameter can be obtained by replacing @xmath90 in eqs .
( [ i1reld ] ) and ( [ i1eins ] ) .
in fact , we will henceforth treat both models a and b simultaneously by setting @xmath91 , where @xmath92 and @xmath93 respectively represent the nonconserved and conserved cases .
explicitly , we thus obtain @xmath94}{\delta s^\alpha(\vec{x},t ) } + \zeta^\alpha(\vec{x},t ) \nonumber \\ & = & - d \ ,
( \i \vec{\nabla})^a \bigl [ r - \vec{\nabla}^2 + \frac{u}{6 } \sum_\beta [ s^\beta(\vec{x})]^2 \bigr ] s^\alpha(\vec{x},t ) \nonumber \\ & & + d \ , ( \i \vec{\nabla})^a \ , h^\alpha(\vec{x},t ) + \zeta^\alpha(\vec{x},t ) \ , \label{i1lane}\end{aligned}\ ] ] with @xmath95 notice already that the presence or absence of a conservation law for the order parameter implies different dynamics for systems described by identical static behaviour . before proceeding with the analysis of the relaxational models
, we remark that in general there may exist additional _
reversible _ contributions to the systematic forces @xmath96 $ ] , see sec .
[ ssec : i6 ] , and / or dynamical mode - couplings to additional conserved , slow fields , which effect further splitting into several distinct _
dynamic universality classes _ @xcite .
let us now evaluate the dynamic response and correlation functions in the _ gaussian _ ( mean - field ) approximation in the high - temperature phase . to this end , we set @xmath97 and thus discard the nonlinear terms in the hamiltonian ( [ i1glwh ] ) as well as in eq .
( [ i1lane ] ) .
the ensuing langevin equation becomes linear in the fields @xmath56 , and is therefore readily solved by means of fourier transforms .
straightforward algebra and regrouping some terms yields @xmath98 s^\alpha(\vec{q},\omega ) = d \vec{q}^a
\ , h^\alpha(\vec{q},\omega ) + \zeta^\alpha(\vec{q},\omega ) \ . \label{i1gaud}\ ] ] with @xmath99 , this gives immediately @xmath100 with the _ response propagator _
@xmath101^{-1 } \ .
\label{i1resp}\ ] ] as is readily established by means of the residue theorem , its fourier backtransform in time obeys _ causality _ , @xmath102 setting @xmath103 , and with the noise correlator ( [ i1ncor ] ) in fourier space @xmath104 we obtain the gaussian dynamic correlation function @xmath105 , where @xmath106 ^ 2 } = 2 k_{\rm b}t \ , d \vec{q}^a \
, |g_0(\vec{q},\omega)|^2 \ . \label{i1gcor}\ ] ] the fluctuation dissipation theorem ( [ i1corw ] ) is of course satisfied ; moreover , as function of wavevector and time , @xmath107 in the gaussian approximation , away from criticality ( @xmath108 , @xmath109 ) the temporal correlations for models a and b decay exponentially , with the relaxation rate @xmath110 . upon comparison with the dynamic scaling hypothesis ( [ i1dysc ] )
, we infer the mean - field scaling exponents @xmath111 and @xmath112 . at the critical point
, a nonconserved order parameter relaxes diffusively ( @xmath113 ) in this approximation , whereas the conserved order parameter kinetics becomes even slower , namely subdiffusive with @xmath114 . finally , invoking eqs .
( [ i1susc ] ) , ( [ i1corw ] ) , ( [ i1cort ] ) , or simply the static limit @xmath115 , we find @xmath116 for the gaussian model
. the full nonlinear langevin equation ( [ i1lane ] ) can not be solved exactly .
yet a perturbation expansion with respect to the coupling @xmath117 may be set up in a slightly cumbersome , but straightforward manner by direct iteration of the equations of motion @xcite .
more elegantly , one may utilise a path - integral representation of the langevin stochastic process @xcite , which allows the application of all the standard tools from statistical and quantum field theory @xcite , and has the additional advantage of rendering symmetries in the problem more explicit @xcite .
our starting point is a set of coupled langevin equations of the form ( [ i1land ] ) for mesoscopic , coarse - grained stochastic variables @xmath80 .
for the stochastic forces , we make the simplest possible assumption of _ gaussian white noise _ , @xmath118 where @xmath119 may represent a differential operator ( such as the laplacian @xmath120 for conserved fields ) , and even a functional of @xmath56 . in the time interval @xmath121 , the moments ( [ i2noic ] ) are encoded in the probability distribution @xmath122 \propto \exp \biggl [ - \frac14 \int \ !
\d^dx \ !
\int_0^{t_f } \ !
\d t \sum_\alpha \zeta^\alpha(\vec{x},t ) \left [ ( l^\alpha)^{-1 } \zeta^\alpha(\vec{x},t ) \right ] \biggr ] \ . \label{i2nprd}\ ] ] if we now switch variables from the stochastic noise @xmath123 to the fields @xmath56 by means of the equations of motion ( [ i1land ] ) , we obtain @xmath122 \ , { \cal d}[\zeta ] = { \cal p}[s ] \ , { \cal d}[s ] \propto \e^{- { \cal g}[s ] } \ , { \cal d}[s ] \ , \label{i2vart}\ ] ] with the statistical weight determined by the _ onsager
machlup functional _
@xcite @xmath124 = \frac14 \int \
! \int \ ! \d t \sum_\alpha \left ( \frac{\partial s^\alpha}{\partial t } - f^\alpha[s ] \right ) \left [ ( l^\alpha)^{-1 } \left ( \frac{\partial s^\alpha}{\partial t } - f^\alpha[s ] \right ) \right ] \ . \label{i2onma}\ ] ] note that the jacobian for the nonlinear variable transformation @xmath125 has been omitted here .
in fact , the above procedure is properly defined through appropriately discretising time . if a _ forward _ ( it ) discretisation is applied , then indeed the associated functional determinant is a mere constant that can be absorbed in the functional measure .
the functional ( [ i2onma ] ) already represents a desired field theory action .
since the probability distribution for the stochastic forces should be normalised , @xmath126 \ , w[\zeta ] = 1 $ ] , the associated ` partition function ' is unity , and carries no physical information ( as opposed to static statistical field theory , where it determines the free energy and hence the entire thermodynamics ) .
the onsager
machlup representation is however plagued by technical problems : eq .
( [ i2onma ] ) contains @xmath127 , which for conserved variables entails the inverse laplacian operator , i.e. , a green function in real space or the singular factor @xmath128 in fourier space ; moreover the nonlinearities in @xmath96 $ ] appear _
hence it is desirable to linearise the action ( [ i2onma ] ) by means of a hubbard stratonovich transformation @xcite .
we shall follow an alternative , more general route that completely avoids the appearance of the inverse operators @xmath127 in intermediate steps .
our goal is to average over noise ` histories ' for observables @xmath129 $ ] that need to be expressible in terms of the stochastic fields @xmath56 : @xmath130 \rangle_\zeta \propto \int \ ! { \cal d}[\zeta ] \ , a[s(\zeta ) ] \ , w[\zeta]$ ] .
for this purpose , we employ the identity @xmath131 \prod_\alpha \prod_{(\vec{x},t ) } \ , \delta\left ( \frac{\partial s^\alpha(\vec{x},t)}{\partial t } - f^\alpha[s](\vec{x},t ) - \zeta^\alpha(\vec{x},t ) \right ) \nonumber \\ & = & \int \ ! { \cal d}[\i { \widetilde s } ] \ ! \int \ ! { \cal d}[s ] \exp \biggl [ - \int \
\d t \sum_\alpha { \widetilde s}^\alpha \left ( \frac{\partial s^\alpha}{\partial t } - f^\alpha[s ] - \zeta^\alpha \right ) \biggr ] \
, \quad \label{i2fcid}\end{aligned}\ ] ] where the first line constitutes a rather involved representation of the unity ( in a somewhat symbolic notation ; again proper discretisation should be invoked here ) , and the second line utilises the fourier representation of the ( functional ) delta distribution by means of the purely imaginary auxiliary fields @xmath132 ( and factors @xmath133 have been absorbed in its functional measure ) . inserting ( [ i2fcid ] ) and the probability distribution ( [ i2nprd ] ) into the desired stochastic noise average , we arrive at @xmath134 \rangle_\zeta & \propto & \int \ !
{ \cal d}[\i { \widetilde s } ] \ !
\int \ ! { \cal d}[s ] \exp \ ! \biggl [ - \ ! \int \ ! \d^dx \
\d t \sum_\alpha { \widetilde s}^\alpha \left(\frac{\partial s^\alpha}{\partial t } - f^\alpha[s ] \right ) \biggr ] \ , a[s ] \nonumber \\ & \times & \int \ ! { \cal d}[\zeta ] \exp \biggl ( - \int \ ! \d^dx \ ! \int \ ! \d t \sum_\alpha \biggl [ \ , \frac14 \ , \zeta^\alpha ( l^\alpha)^{-1 } \zeta^\alpha - { \widetilde s}^\alpha \ , \zeta^\alpha \biggr ] \biggr ) \ . \label{i2noia}\end{aligned}\ ] ] we may now evaluate the gaussian integrals over the noise @xmath123 , which yields @xmath135 \rangle_\zeta = \int \ !
{ \cal d}[s ] \ , a[s ] \ , { \cal p}[s ] \ , \quad { \cal p}[s ] \propto \int \ ! { \cal d}[\i { \widetilde s } ] \ , \e^{- { \cal a}[{\widetilde s},s ] } \ , \label{i2noip}\ ] ] with the statistical weight now governed by the _ janssen de dominicis ` response ' functional
_ @xcite @xmath136 = \int \ ! \d^dx \ !
\int_0^{t_f } \ !
\d t \sum_\alpha \left [ { \widetilde s}^\alpha \left ( \frac{\partial s^\alpha}{\partial t } - f^\alpha[s ] \right ) - { \widetilde s}^\alpha \ , l^\alpha \ , { \widetilde s}^\alpha \right ] \ . \label{i2jade}\ ] ] once again , we have omitted the functional determinant from the variable change @xmath125 , and normalisation implies @xmath137 \int \ ! { \cal d}[s ] \ ,
\e^{- { \cal a}[{\widetilde s},s ] } = 1 $ ] . the first term in the action ( [ i2jade ] ) encodes the temporal evolution according to the systematic terms in the langevin equations ( [ i1land ] ) , whereas the second term specifies the noise correlations ( [ i2noic ] ) . since the auxiliary variables @xmath138 , often termed martin siggia
rose response fields @xcite , appear only quadratically here , they may be eliminated via completing the squares and gaussian integrations ; thereby one recovers the onsager
machlup functional ( [ i2onma ] ) . the janssen
de dominicis functional ( [ i2jade ] ) takes the form of a ( @xmath139)-dimensional statistical field theory with two independent sets of fields @xmath56 and @xmath138 .
we may thus bring the established machinery of statistical and quantum field theory @xcite to bear here ; it should however be noted that the response functional formalism for stochastic langevin dynamics incorporates causality in a nontrivial manner , which leads to important distinctions @xcite .
let us specify the janssen
de dominicis functional for the purely _ relaxational models _ a and b @xcite , see eqs .
( [ i1lane ] ) and ( [ i1ncor ] ) , splitting it into the gaussian and anharmonic parts @xmath140 @xcite , which read @xmath141 & = & \int \ !
\d t \sum_\alpha \biggl ( { \widetilde s}^\alpha \left [ \frac{\partial}{\partial t } + d \ , ( \i \vec{\nabla})^a \ , ( r - \vec{\nabla}^2 ) \right ] s^\alpha \nonumber \\ & & \qquad\qquad\qquad\quad - d \ , { \widetilde s}^\alpha \ , ( \i \vec{\nabla})^a \ , { \widetilde s}^\alpha - d \ , { \widetilde s}^\alpha \ , ( \i \vec{\nabla})^a \ , h^\alpha \biggr ) \ , \label{i2hajd } \\ { \cal a}_{\rm int}[{\widetilde s},s ] & = & d \ , \frac{u}{6 } \int \ ! \d^dx \
! \int \ ! \d t \sum_{\alpha,\beta } { \widetilde s}^\alpha \ ,
( \i \vec{\nabla})^a \ , s^\alpha \ , s^\beta \ , s^\beta \ .
\label{i2ndjd}\end{aligned}\ ] ] since we are interested in the vicinity of the critical point @xmath142 , we have absorbed the constant @xmath143 into the fields .
the prescription ( [ i2noip ] ) tells us how to compute time - dependent correlation functions @xmath144 . using eq .
( [ i2hajd ] ) , the dynamic order parameter _ susceptibility
_ follows from @xmath145 for the simple relaxational models ( only ) , the response function is just given by a correlator that involves an auxiliary variable , which explains why the @xmath138 are referred to as ` response ' fields . in equilibrium , one may employ the onsager
machlup functional ( [ i2onma ] ) to derive the _ fluctuation dissipation theorem _
@xcite @xmath146 which is equivalent to eq .
( [ i1corw ] ) in fourier space . in order to access arbitrary correlators
, we define the _ generating functional _ @xmath147 = \biggl\langle \exp \int \ !
! \d t \sum_\alpha \left ( { \tilde j}^\alpha \ , { \widetilde s}^\alpha + j^\alpha \ , s^\alpha \right ) \biggr\rangle \ , \label{i2genf}\ ] ] wherefrom the correlation functions follow via functional derivatives , @xmath148 \bigg\vert_{{\tilde j}=0=j } \ , \label{i2corf}\ ] ] and the _ cumulants _ or _
connected _ correlation functions via @xmath149 \bigg\vert_{{\tilde j}=0=j } \ .
\label{i2cumf}\ ] ] in the harmonic approximation , setting @xmath97 , @xmath150 $ ] can be evaluated explicitly ( most directly in fourier space ) by means of gaussian integration @xcite ; one thereby recovers ( with @xmath151 ) the gaussian response propagator ( [ i1resp ] ) and two - point correlation function ( [ i1gcor ] ) .
moreover , as a consequence of causality , @xmath152 . since we can not evaluate correlation functions with the nonlinear action ( [ i2ndjd ] ) exactly , we resort to a perturbational treatment , assuming , for the time being , a small coupling strength @xmath117 .
perturbation expansion _ with respect to @xmath117 is constructed by rewriting the desired correlation functions in terms of averages with respect to the gaussian action ( [ i2hajd ] ) , henceforth indicated with index @xmath153 , and then expanding the exponential of @xmath154 , @xmath155 } \right\rangle_0}{\left\langle e^{-{\cal a}_{\rm int}[{\widetilde s},s ] } \right\rangle_0 } \nonumber \\ & = & \biggl\langle \prod_{ij } s^{\alpha_i } \ , { \widetilde s}^{\alpha_j } \sum_{l=0}^\infty \frac{1}{l ! } \left ( -{\cal a}_{\rm int}[{\widetilde s},s ] \right)^l \biggr\rangle_0 \ . \label{i3pexp}\end{aligned}\ ] ] the remaining gaussian averages , a series of polynomials in the fields @xmath56 and @xmath138 , can be evaluated by means of _ wick s theorem _
, here an immediate consequence of the gaussian statistical weight , which states that all such averages can be written as a sum over all possible factorisations into gaussian two - point functions @xmath156 , i.e. , essentially the response propagator @xmath157 , eq .
( [ i1resp ] ) , and @xmath158 , the gaussian correlation function @xmath159 , eq .
( [ i1gcor ] ) . recall that the denominator in eq .
( [ i3pexp ] ) is exactly unity as a consequence of normalisation ; alternatively , this result follows from causality in conjunction with our forward descretisation prescription , which implies that we should identify @xmath160 .
( we remark that had we chosen another temporal discretisation rule , any apparent contributions from the denominator would be precisely cancelled by the in this case nonvanishing functional jacobian from the variable transformation @xmath125 . ) at any rate , our stochastic field theory contains _ no ` vacuum ' contributions_. the many terms in the perturbation expansion ( [ i3pexp ] ) are most lucidly organised in a graphical representation , using _
feynman diagrams _ with the basic elements depicted in fig .
[ relgra ] .
we represent the response propagator ( [ i1resp ] ) by a _ directed line _ ( here conventionally from right to left ) , which encodes its causal nature ; the noise by a two - point _ ` source ' vertex _ , and the anharmonic term in eq .
( [ i2ndjd ] ) as a _ four - point vertex_. in the diagrams representing the different terms in the perturbation series , these vertices serve as links for the propagator lines , with the fields @xmath56 being encoded as the ` incoming ' , and the @xmath138 as the ` outgoing ' components of the lines . in fourier space , translational
invariance in space and time implies wavevector and frequency conservation at each vertex , see fig .
[ temver ] below .
an alternative , equivalent representation uses both the response and correlation propagators as independent elements , the latter depicted as undirected line , thereby disposing of the noise vertex , and retaining the nonlinearity in fig . [
relgra](c ) as sole vertex .
following standard field theory procedures @xcite , one establishes that the perturbation series for the _ cumulants _ ( [ i2cumf ] ) is given in terms of _ connected _ feynman graphs only ( for a detailed exposition of this and the following results , see ref .
@xcite ) . an additional helpful reduction in the number of diagrams to be considered arises when one considers the _ vertex functions _ , which generalise the self - energy contributions @xmath161 in the _ dyson equation _ for the response propagator , @xmath162 . to this end
, we define the fields @xmath163 and @xmath164 , and introduce the new _ generating functional _ @xmath165 = - \ln { \cal z}[{\tilde j},j ] + \int \ ! \d^dx \ ! \int \ ! \d t \sum_\alpha \left ( { \tilde j}^\alpha \ , { \widetilde \phi}^\alpha + j^\alpha \ , \phi^\alpha \right ) \ , \label{i3vgen}\ ] ] wherefrom the vertex functions are obtained via the functional derivatives @xmath166 \bigg\vert_{{\tilde j}=0=j } \ . \label{i3verf}\ ] ] diagrammatically , these quantities turn out to be represented by the possible sets of _ one - particle ( 1pi ) irreducible feynman graphs _ with @xmath167 incoming and @xmath168 outgoing ` amputated ' legs ; i.e. , these diagrams do not split into allowed subgraphs by simply cutting any single propagator line .
for example , for the two - point functions a direct calculation yields the relations @xmath169 where the second equation for @xmath170 follows from the fluctuation
dissipation theorem ( [ i1corw ] ) .
note that @xmath171 vanishes because of causality .
the perturbation series can then be organised graphically as an expansion in successive orders with respect to the number of closed propagator _ loops_. as an example , fig .
[ temver ] depicts the one - loop contributions for the vertex functions @xmath172 and @xmath173 in the time domain with all required labels .
one may formulate general _ feynman rules _ for the construction of the diagrams and their translation into mathematical expressions for the @xmath174th order contribution to the _ vertex function _ @xmath175 : 1 . draw all topologically different , connected _ one - particle irreducible graphs _ with @xmath168 outgoing and @xmath167 incoming lines connecting @xmath174 relaxation vertices @xmath176 .
do _ not _ allow closed response loops ( since in the it calculus @xmath160 ) .
2 . attach wavevectors @xmath177 , frequencies @xmath178 or times @xmath179 , and component indices @xmath180 to all directed lines , obeying ` momentum ( and energy ) ' conservation at each vertex .
3 . each directed line corresponds to a response propagator @xmath181 or @xmath182 in the frequency and time domain , respectively , the two - point vertex to the noise strength @xmath183 , and the four - point relaxation vertex to @xmath184 .
closed loops imply integrals over the internal wavevectors and frequencies or times , subject to causality constraints , as well as sums over the internal vector indices .
apply the residue theorem to evaluate frequency integrals .
4 . multiply with @xmath185 and the combinatorial factor counting all possible ways of connecting the propagators , @xmath174 relaxation vertices , and @xmath186 two - point vertices leading to topologically identical graphs , including a factor @xmath187 originating in the expansion of @xmath188)$ ] . for later use
, we provide the explicit results for the two - point vertex functions to two - loop order . after some algebra , the three diagrams in fig .
[ renpro ] give @xmath189 \
, \label{i3ga11}\end{aligned}\ ] ] where we have separated out the dynamic part in the last line , and introduced the abbreviations @xmath190 and @xmath191 @xcite . for the noise vertex , fig .
[ renver](a ) yields @xcite @xmath192 \ ; \quad \label{i3ga20}\end{aligned}\ ] ] notice that for model b , as a consequence of the conservation law for the order parameter and ensuing wavevector dependence of the nonlinear vertex , see fig .
[ relgra](c ) , to _ all orders _ in the perturbation expansion @xmath193 at last , with the shorthand notation @xmath194 , the analytical expression corresponding to the graph in fig .
[ renver](b ) for the four - point vertex function at symmetrically chosen external wavevector labels reads @xmath195 \ . \label{i3ga13}\end{aligned}\ ] ] consider a typical loop integral , say the correction in eq .
( [ i3ga13 ] ) to the four - point vertex function @xmath173 at zero external frequency and momentum , whose ` bare ' value , without any fluctuation contributions , is @xmath117 . in dimensions
@xmath196 , one obtains , after introducing @xmath13-dimensional spherical coordinates and rendering the integrand dimensionless ( @xmath197 ) : @xmath198 where we have inserted the surface area @xmath199 of the @xmath13-dimensional unit sphere , with euler s gamma function , @xmath200 .
note that the integral on the right - hand side is finite .
thus , we see that the _ effective _ expansion parameter in perturbation theory is not just @xmath117 , but the combinaton @xmath201 . far away from @xmath4 ,
it is small , and the perturbation expansion well - defined .
however , @xmath202 as @xmath203 for @xmath196 : we are facing _ infrared ( ir ) divergences _ , induced by the strong critical fluctuations that render the loop corrections singular
. a straightforward application of perturbation theory will therefore not provide meaningful results , and we must expect the fluctuation contributions to modify the critical power laws .
conversely , for dimensions @xmath204 , the integral in ( [ i4flin ] ) develops _ ultraviolet ( uv ) divergences _ as the upper integral boundary is sent to infinity ( @xmath205 ) , @xmath206 in lattice models , there is a finite wavevector cutoff , namely the brillouin zone boundary , @xmath207 for a hypercubic lattice with lattice constant @xmath208 , whence physically these uv problems do not emerge .
yet we shall see that a formal treatment of these unphysical uv divergences will allow us to infer the correct power laws for the physical ir singularities associated with the critical point .
the borderline dimension that separates the ir and uv singular regimes is referred to as _ upper critical dimension _
@xmath0 ; here @xmath209 .
note that at @xmath0 , uv and ir singularities are intimately connected and appear in the form of _ logarithmic divergences _
, see eq .
( [ i4uvsi ] ) .
the situation is summarised in table [ iruvst ] , where we have also stated that models with continuous order parameter symmetry , such as the hamiltonian ( [ i1glwh ] ) with @xmath210 , do not allow long - range order in dimensions @xmath211 ( _ mermin wagner hohenberg _ theorem @xcite ) . here
, @xmath212 is called the _ lower critical dimension _ ; for the ising model represented by eq .
( [ i1glwh ] ) with @xmath213 , of course @xmath214 .
llll dimension & perturbation & model a / b or & critical + interval & series & @xmath215 field theory & behaviour + @xmath216 & ir - singular & ill - defined & no long - range + & uv - convergent & @xmath117 relevant & order ( @xmath210 ) + @xmath217 & ir - singular & super - renormalisable & nonclassical + & uv - convergent & @xmath117 relevant & exponents + @xmath218 & logarithmic ir-/ & renormalisable & logarithmic + & uv - divergence & @xmath117 marginal & corrections + @xmath219 & ir - regular & nonrenormalisable & mean - field + & uv - divergent & @xmath117 irrelevant & exponents + the upper critical dimension can be obtained in a more direct manner through simple _ power counting_. to this end , we introduce an arbitrary momentum scale @xmath220 , i.e. , define the _ scaling dimensions _ @xmath221 = \mu^{-1}$ ] and @xmath222 = \mu$ ] . if in addition we choose @xmath223 = \mu^{-2-a}$ ] , or @xmath224 = \mu^{2+a}$ ] , then the relaxation constant becomes dimensionless , @xmath225 = \mu^0 $ ] . for the deviation from the critical point ,
we obtain @xmath226 = \mu^2 $ ] , and the _ positive _ exponent indicates that this control parameter constitutes a _ relevant _ coupling in the theory ; as we shall see below , its renormalised counterpart grows under subsequent rg transformations . for the nonlinear coupling ,
one finds @xmath227 = \mu^{4-d}$ ] , so it is relevant for @xmath196 : nonlinear thermal fluctuations will qualitatively affect the physical properties at the phase transition ; but @xmath117 becomes irrelevant for @xmath219 : one then expects mean - field ( gaussian ) critical exponents . at the upper critical dimension @xmath209 , the nonlinear coupling @xmath117 is _ marginally relevant _ : this will induce logarithmic corrections to the mean - field scaling laws , see table [ iruvst ] .
it is obviously not a simple task to treat the ir - singular perturbation expansion in a meaningful , well - defined manner , and thus allow nonanalytic modifications of the critical power laws ( note that mean - field scaling is completely determined by dimensional analysis or power counting ) .
the key of the success of the rg approach is to focus on the very specific symmetry that emerges near critical points , namely _
scale invariance_. there are several ( largely equivalent ) versions of the rg method ; we shall here formulate and employ the field - theoretic variant @xcite . in order to proceed , it is convenient to evaluate the loop integrals in momentum space by means of _ dimensional regularisation _ , whereby one assigns finite values even to uv - divergent expressions , namely the analytically continued values from the uv - finite range .
for example , even for noninteger dimensions @xmath13 and @xmath228 , we set @xmath229 the _ renormalisation _ program then consists of the following steps : 1 .
we aim to carefully keep track of formal , unphysical uv divergences . in
dimensionally regularised integrals ( [ i4dimr ] ) , these appear as poles in @xmath230 ; their residues characterise the asymptotic uv behaviour of the field theory under consideration . 2 .
therefrom we may infer the ( uv ) scaling properties of the control parameters of the model under a rg transformation , namely essentially a change of the momentum scale @xmath220 , while keeping the form of the action invariant .
this will allow us to define suitable _ running couplings_. 3 .
we seek _ fixed points _ in parameter space where certain marginal couplings ( @xmath117 here ) do not change anymore under rg transformations .
this describes a _ scale - invariant _
regime for the model under consideration , where the uv and ir scaling properties become intimately linked . studying
the parameter flows near a stable rg fixed point then allows us to extract the asymptotic ir power laws . as a preliminary step , we need to take into account that the fluctuations will also shift the critical point downwards from the mean - field phase transition temperature @xmath54 ; i.e. , we expect the transition to occur at @xmath231
. this fluctuation - induced @xmath4 shift can be determined by demanding that the inverse static susceptibility vanish at @xmath4 : @xmath232 , where @xmath233 and thus @xmath234 .
using our previous results ( [ i3gm11 ] ) and ( [ i3ga11 ] ) , we find to first order in @xmath117 ( and with finite cutoff @xmath235 ) , @xmath236 notice that this quantity depends on microscopic details ( the lattice structure enters the cutoff @xmath235 ) and is thus not universal ; moreover it diverges for @xmath237 ( quadratically near @xmath209 ) as @xmath238 .
we next use @xmath239 to write physical quantities as functions of the true distance @xmath71 from the critical point , which technically amounts to an _ additive renormalisation _ ; e.g. , the dynamic response function becomes to one - loop order @xmath240 + o(u^2 ) \ .
\label{i4dsus}\ ] ] the remaining loop integral is uv - singular in dimensions @xmath241 .
we may now formally absorb the remaining uv divergences into _ renormalised _ fields and parameters , a procedure called _ multiplicative renormalisation_. for the renormalised fields , we use the convention @xmath242 where we have exploited the @xmath46 rotational symmetry in using identical _ renormalisation constants ( @xmath243 factors ) _ for each component .
the renormalised cumulants with @xmath167 order parameter fields @xmath56 and @xmath168 response fields @xmath138 naturally involve the product @xmath244 , whence @xmath245 in a similar manner , we relate the ` bare ' parameters of the theory via @xmath243 factors to their renormalised counterparts , which we furthermore render dimensionless through appropriate momentum scale factors , @xmath246 where we have separated out the factor @xmath247 for convenience . in the _ minimal subtraction _ scheme ,
the @xmath243 factors contain _ only _ the uv - singular terms , which in dimensional regularisation appear as poles at @xmath248 , and their residues , evaluated at @xmath249 .
these renormalisation constants are not all independent , however ; since the equilibrium fluctuation dissipation theorem ( [ i2fldt ] ) or ( [ i1corw ] ) must hold in the renormalised theory as well , we infer that necessarily @xmath250 and consequently from eq .
( [ i3gm11 ] ) @xmath251 moreover , for model b with conserved order parameter eq .
( [ i3mbvf ] ) implies that to all orders in the perturbation expansion @xmath252 for the following , it is crucial that the theory is _
renormalisable _ , i.e. , a _ finite _ number of reparametrisations suffice to formally rid it of all uv divergences . indeed , for the relaxational models a and b , and the static ginzburg
wilson hamiltonian ( [ i1glwh ] ) , all higher vertex function beyond the four - point function are uv - convergent near @xmath0 , and there are only the _ three _ independent static renormalisation factors @xmath253 , @xmath254 , and @xmath255 , and in addition @xmath256 for nonconserved order parameter dynamics . as we shall see , these directly translate into the _ two _ independent static critical exponents and the unrelated dynamic scaling exponent @xmath60 for model a ; for model b with conserved order parameter , eq .
( [ i4brel ] ) will yield a scaling relation between @xmath60 and @xmath14 . in order to explicitly determine the renormalisation constants , we need to ensure that we stay away from the ir - singular regime .
this is guaranteed by selecting as _ normalisation point _
either @xmath257 ( i.e. , @xmath258 ) or @xmath259 . inevitably therefore , the renormalised theory depends on the corresponding arbitrary momentum scale @xmath220 .
since there are no fluctuation contributions to order @xmath117 to either @xmath260 or @xmath261 ( at @xmath257 ) , we find @xmath262 and @xmath263 within the one - loop approximation . expressions ( [ i4dsus ] ) and ( [ i3ga13 ] ) then yield with the formula ( [ i4dimr ] ) @xmath264 to two - loop order , we may infer the field renormalisation @xmath253 from the static susceptibility as the singular contributions to @xmath265 , and @xmath256 for model a , through a somewhat lengthy calculation @xcite , from either @xmath266 or @xmath267 , with the results @xmath268 we now wish to related the renormalised vertex functions at different inverse length scales @xmath220 .
this is accomplished by simply recalling that the _ unrenormalised _ vertex functions obviously do _ not _ depend on @xmath220 , @xmath269 \ . \label{i5unrm}\ ] ] in the second step , the bare quantities have been replaced with their renormalised counterparts . the innocuous statement ( [ i5unrm ] ) then implies a very nontrivial partial differential equation for the renormalised vertex functions , the desired _ renormalisation group equation _ , @xmath270 \nonumber \\ & & \qquad \times \ \gamma_r^{({\widetilde n},n)}\left ( \mu , d_r,\tau_r , u_r \right ) = 0 \ .
\label{i5rgeq}\end{aligned}\ ] ] here we have defined _ wilson s flow functions _
( the index ` 0 ' indicates that the derivatives with respect to @xmath220 are to be taken with fixed unrenormalised parameters ) @xmath271 where we have used the relation ( [ i4zrel ] ) ; for model b , eq .
( [ i4brel ] ) gives in addition @xmath272 we have also introduced the _
rg beta function _ for the nonlinear coupling @xmath117 , @xmath273 explicitly , eqs .
( [ i42lpz ] ) and ( [ i41lpz ] ) yield to lowest nontrivial order , with @xmath274 , @xmath275 \ .
\label{i5betu}\end{aligned}\ ] ] in the rg equation for the renormalised dynamic susceptibility , eq .
( [ i4cren ] ) tells us that the second term in eq .
( [ i5rgeq ] ) is to be replaced with @xmath276 .
its explicit dependence on the scale @xmath220 can be factored out via @xmath277 , see eq .
( [ i4dsus ] ) , whence @xmath278 { \hat \chi}_r(d_r,\tau_r , u_r ) = 0 \ .
\label{i5rgsu}\ ] ] this _ linear partial differential equation _ is readily solved by means of the _ method of characteristics _ , as is eq .
( [ i5rgeq ] ) for the vertex functions .
the idea is to find a curve parametrisation @xmath279 in the space spanned by the parameters @xmath280 , @xmath281 , and @xmath282 such that @xmath283 with initial values @xmath284 , @xmath285 , and @xmath286 , respectively at @xmath287 . the _ first - order ordinary differential equations _ ( [ i5runc ] ) , with @xmath288 etc . define _ running couplings _ that describe how the parameters of the theory change under scale transformations @xmath289 .
the formal solutions for the @xmath290 and @xmath290 read @xmath291 \
, \quad { \widetilde \tau}(\ell ) = \tau_r \ , \exp \left [ \int_1^\ell \ ! \gamma_\tau(\ell ' ) \ , \frac{\d \ell'}{\ell ' } \right ] \ . \label{i5irun}\ ] ] for the function @xmath292 , we then obtain another ordinary differential equation , namely @xmath293 { \hat \chi}(\ell ) \ , \label{i5odfs}\ ] ] which is solved by @xmath294 \ .
\label{i5rgsl}\ ] ] collecting everything , we finally arrive at @xmath295 \nonumber \\ & & \times { \hat \chi}_r\left ( { \widetilde \tau}(\ell ) , { \widetilde u}(\ell ) , \frac{|\vec{q}|}{\mu \ , \ell } , \frac{\omega}{{\widetilde d}(\ell ) \ , ( \mu \ , \ell)^{2+a } } \right ) \ . \label{i5rchi}\end{aligned}\ ] ] the solution ( [ i5rchi ] ) of the rg equation ( [ i5rgsu ] ) , along with the flow equations ( [ i5runc ] ) , ( [ i5irun ] ) for the running couplings tell us how the dynamic susceptibility depends on the ( momentum ) scale @xmath296 at which we consider the theory .
similar relations can be obtained for arbitrary vertex functions by solving the associated rg equations ( [ i5rgeq ] ) @xcite .
the point here is that the right - hand side of eq .
( [ i5rchi ] ) may be evaluated outside the ir - singular regime , by fixing one of its arguments at a finite value , say @xmath297 .
the function @xmath298 is regular , and can be calculated by means of perturbation theory .
scale - invariant _ regime is characterised by the renormalised nonlinear coupling @xmath286 becoming independent of the scale @xmath296 , or @xmath299 . for an _ rg fixed point _ to be _ infrared - stable _
, we thus require @xmath300 since eq .
( [ i5runc ] ) then implies that @xmath301 . taking the limit @xmath302 thus provides the desired mapping of physical observables such as ( [ i5rchi ] ) onto the critical region . in the vicinity of an ir - stable rg fixed point , eq .
( [ i5irun ] ) yields the power laws @xmath303 , where @xmath304 , etc . consequently , eq .
( [ i5rchi ] ) reduces to @xmath305 and upon _ matching _
@xmath306 we recover the _
dynamic scaling law _ ( [ i1susc ] ) with the _ critical exponents _
@xmath307 to one - loop order , we obtain from the rg beta function ( [ i5betu ] ) @xmath308 here we have indicated that our perturbative expansion for small @xmath117 has effectively turned into a _ dimensional expansion _ in @xmath230 . in dimensions
@xmath196 , the _ heisenberg _ fixed point @xmath309 is ir - stable , since @xmath310 . with eqs . ( [ i5gams ] ) and ( [ i5gamt ] )
, the identifications ( [ i5crex ] ) then give us explicit results for the static scaling exponents , as mere functions of dimension @xmath311 and the number of order parameter components @xmath312 , @xmath313 for model a with nonconserved order parameter , the two - loop result ( [ i5gamd ] ) yields the independent dynamic critical exponent @xmath314 for model b with conserved order parameter , instead @xmath315 , whence we arrive at the _ exact _ scaling relation @xmath316 in dimensions @xmath317 , the _ gaussian _
fixed point @xmath318 is stable ( @xmath319 ) .
therefore all anomalous dimensions disappear , i.e. , @xmath320 and @xmath321 , and we are left with the mean - field critical exponents @xmath116 , @xmath111 , and @xmath112 . precisely at the upper critical dimension @xmath209 , the rg flow equation for the nonlinear coupling becomes @xmath322 which is solved by @xmath323 in four dimensions , @xmath324 , but only logarithmically slowly , which causes _ logarithmic corrections _ to the mean - field critical power laws .
for example , upon inserting eq .
( [ i5lgus ] ) into the flow equation ( [ i5runc ] ) , one finds @xmath325 ; with @xmath326 , iterative inversion yields @xmath327 this concludes our derivation of asymptotic scaling laws for the critical dynamics of the purely relaxational models a and b , and the explicit computation of the scaling exponents in powers of @xmath230 . in the following sections ,
i will briefly sketch how the response functional formalism and the dynamic renormalisation group can be employed to study the critical dynamics of systems with reversible mode - coupling terms , the ` ageing ' behaviour induced by quenching from random initial conditions to the critical point , the effects of violating the detailed balance constraints on universal dynamic critical properties , and the generically scale - invariant features of nonequilibrium systems such as driven diffusive ising lattice gases . in the previous chapters ,
we have assumed purely relaxational dynamics for the order parameter , see eq .
( [ i1reld ] ) . in general , however , there are also _ reversible _ contributions to the systematic force terms @xmath328 that enter its langevin equation @xcite .
consider the hamiltonian dynamics of _ microscopic _ variables , say , local spin densities , at @xmath329 : @xmath330,s_m^\alpha(x , t ) \bigr\}$ ] . here ,
the _ poisson brackets _ @xmath331 constitute the classical analog of the quantum - mechanical commutator @xmath332 $ ] ( correspondence principle ) . upon _ coarse - graining _ , the microscopic variables @xmath333 become the _ mesoscopic _ hydrodynamic fields @xmath56 . since the set of _ slow _ modes
should provide a complete description of the critical dynamics , we may formally expand @xmath334 , s^\alpha(\vec{x } ) \bigr\ } = \int \ ! \d^dx ' \ , \sum_\beta \frac{\delta { \cal h}[s]}{\delta s^\beta(\vec{x } ' ) } \ , q^{\beta \alpha}(\vec{x}',\vec{x } ) \
, \label{i6come}\ ] ] with the mutual poisson brackets of the hydrodynamic variables @xmath335 by inspection of the associated fokker
planck equation , one may then establish an additional _ equilibrium condition _ in order for the time - dependent probability distribution to reach the canonical limit ( [ i1eqpd ] ) : @xmath86 \to { \cal p}_{\rm eq}[s]$ ] as @xmath336 _ provided _ the probability current is _ divergence - free _ in the space spanned by the stochastic fields @xmath47 : @xmath337 \ e^{- { \cal h}[s ] / k_{\rm b } t } \right ) = 0 \ . \label{i6divf}\ ] ] it turns out that this equilibrium condition is often more crucial than the einstein relation ( [ i1eins ] ) . in order to satisfy eq .
( [ i6divf ] ) at @xmath338 , we must supplement eq .
( [ i6come ] ) by a finite - temperature correction , whereupon the _
reversible mode - coupling _ contributions to the systematic forces become @xmath339(\vec{x } ) = - \int \ !
\d^dx ' \sum_\beta \left [ q^{\alpha \beta}(\vec{x},\vec{x } ' ) \frac{\delta { \cal h}[s]}{\delta s^\beta(\vec{x } ' ) } - k_{\rm b } t \ ,
\frac{\delta q^{\alpha \beta}(\vec{x},\vec{x}')}{\delta s^\beta(\vec{x } ' ) } \right ] \ , \label{i6revm}\ ] ] and the complete coupled set of stochastic differential equations reads @xmath340(\vec{x},t ) - d^\alpha ( \i \vec{\nabla})^{a_\alpha } \frac{\delta { \cal h}[s]}{\delta s^\alpha(\vec{x},t ) } + \zeta^\alpha(\vec{x},t ) \ , \label{i6lang}\ ] ] where as before the @xmath341 denote the relaxation coefficients , and @xmath342 or @xmath343 respectively for nonconserved and conserved modes . as an instructive example , let us consider the _
heisenberg model _ for _ isotropic ferromagnets _
, @xmath344 = - \frac{1}{2 } \sum_{j , k=1}^n j_{jk } \ , \vec{s}_j \cdot \vec{s}_k$ ] ,
where the spin operators satisfy the usual commutation relations @xmath345 = i \hbar \sum_\gamma \epsilon^{\alpha \beta \gamma } s_j^\gamma \ , \delta_{jk}$ ] .
the corresponding poisson brackets for the magnetisation density read @xmath346 where the purely _ dynamical coupling _
@xmath347 incorporates various factors that emerge upon coarse - graining and taking the continuum limit .
the second contribution in eq .
( [ i6revm ] ) vanishes , since it reduces to a contraction of the antisymmetric tensor @xmath348 with the kronecker symbol @xmath349 , whence we arrive at the langevin equations governing the critical dynamics of the three order parameter components for isotropic ferromagnets @xcite @xmath350}{\delta \vec{s}(\vec{x},t ) } + d \ , \vec{\nabla}^2 \ , \frac{\delta { \cal h}[{\vec s}]}{\delta { \vec s}(\vec{x},t ) } + \vec{\zeta}(\vec{x},t ) \ , \label{i6modj}\ ] ] with @xmath351 . since @xmath352 , \sum_k s_k^\alpha \bigr ] = 0 $ ] , the total magnetisation is conserved , whence the noise correlators should be taken as @xmath353 the vector product term in eq .
( [ i6modj ] ) describes the spin precession in the local effective magnetic field @xmath354/\delta \vec{s}$ ] , which includes a contribution induced by the exchange interaction .
the langevin equation ( [ i6modj ] ) and ( [ i6corj ] ) with the hamiltonian ( [ i1glwh ] ) for @xmath355 define the so - called _ model j _ @xcite .
in addition to the model b response functional ( [ i2hajd ] ) and ( [ i2ndjd ] ) with @xmath93 ( setting @xmath151 again ) , the reversible force in eq .
( [ i6modj ] ) leads to an additional contribution to the action @xmath356 = - g \int \ ! \d^dx \
\d t \sum_{\alpha,\beta,\gamma } \epsilon^{\alpha \beta \gamma } { \widetilde s}^\alpha s^\beta \left ( \nabla^2 s^\gamma + h^\gamma \right)\ , \label{i6mcpl}\ ] ] which gives rise to an additional _ mode - coupling vertex _ , as depicted in fig .
[ modelj](a ) .
power counting yields the scaling dimension @xmath357 = \mu^{3-d/2}$ ] for the associated coupling strength , whence we expect a _ dynamical _ upper critical dimension @xmath358 .
however , since we are investigating a system in thermal equilibrium , we can treat its thermodynamics and static properties separately from its dynamics .
obviously therefore , the static critical exponents must still be given ( to lowest nontrivial order and for @xmath359 ) by eqs .
( [ i5stex ] ) for the three - component heisenberg model with @xmath360 rotational symmetry .
therefore our sole task is to find the dynamic critical exponent @xmath60 .
remarkably , @xmath60 is entirely fixed by the symmetries of the problem and can be determined exactly . to this end
, we exploit the fact that the @xmath56 are the generators of the rotation group ; indeed , it follows from eq .
( [ i6mcpl ] ) that applying a time - dependent external field @xmath361 induces a contribution @xmath362 to the average magnetisation . as a consequence ,
we obtain for the _ nonlinear susceptibility _ @xmath363 , @xmath364 an analogous expression must hold after renormalisation as well .
if we define the dimensionless renormalised mode - coupling according to @xmath365 where @xmath366 , eq .
( [ i6nlsu ] ) implies the identity @xcite @xmath367 for the rg beta function associated with the effective coupling entering the loop corrections , we thus infer @xmath368 consequently , at any _ nontrivial _ ir - stable rg fixed point @xmath369 , we have the _ exact _ scaling relation , valid to _ all _ orders in perturbation theory , @xmath370 since the resulting value for the dynamic exponent , @xmath371 in three dimensions , is markedly smaller than the model b mean - field @xmath114 , we conclude that the reversible spin precession kinetics speeds up the order parameter dynamics considerably @xcite .
an explicit one - loop calculation , either for the propagator self - energy @xmath372 , depicted in fig .
[ modelj](b ) , or the noise vertex @xmath373 , shown in fig .
[ modelj](c ) , yields @xcite @xmath374 which along with @xmath375 confirms that there exists a nontrivial mode - coupling rg fixed point @xmath376 where @xmath377 , which is ir - stable for @xmath378 . as @xmath379 for @xmath219 , we indeed recover the mean - field dynamic exponent @xmath114 in @xmath380 dimensions . with the leading singularity thus isolated ,
the regular scaling functions can be computed numerically to high accuracy within a self - consistent one - loop approximation that also goes under the name _ mode - coupling theory_. details of this procedure , an alternative derivation , and many results of mode - coupling theory as applied to the critical dynamics of magnets and comparisons with experimental data can be found in ref .
@xcite .
typically , reversible force terms of the form ( [ i6revm ] ) involve dynamical couplings of the order parameter to _ other _ conserved , slow variables .
in addition , there may also be static couplings to conserved fields in the hamiltonian .
these various possibilities give rise to a range of different _ dynamic universality classes _ for near - equilibrium critical dynamics @xcite .
we shall not pursue these further here ( for a partial account within the field - theoretic rg approach , see ref .
@xcite ) , but instead proceed and now consider nonequilibrium effects .
we begin with a brief discussion of the _ coarsening _ dynamics of systems described by model a / b kinetics that are rapidly _ quenched _ from a disordered state at @xmath381 to the critical point @xmath142 @xcite .
the situation may be modeled as a relaxation from _ gaussian random initial conditions _ ,
i.e. , the probability distribution for the order parameter at @xmath79 can be taken as @xmath382 \propto e^{-{\cal h}_0[s ] } = \exp \left ( - \frac{\delta}{2 } \int \ ! \d^dx \sum_\alpha \left [ s^\alpha(\vec{x},0 ) - a^\alpha(\vec{x } ) \right]^2 \right ) \ , \label{i7init}\ ] ] where the functions @xmath383 specify the most likely initial configurations .
power counting for the parameter @xmath384 gives @xmath385 = \mu^2 $ ] , whence it is a relevant perturbation that will flow to @xmath386 under the rg .
asymptotically , therefore , the system will be governed by sharp _ dirichlet boundary conditions_. whereas the response propagators remains a causal function of the time difference between applied perturbation and effect , @xmath387 , see eq .
( [ i1rest ] ) , time translation invariance is broken by the initial state in the _ dirichlet correlator _ of the gaussian model , @xmath388 away from criticality , i.e. , for @xmath108 and @xmath109 , temporal correlations decay exponentially fast , and the system quickly approaches the stationary equilibrium state . however , as @xmath10 , the equilibration time diverges according to @xmath389 , and the system never reaches thermal equilibrium .
two - time correlation functions will then depend on both times separately , in a specific manner to be addressed below , a phenomenon termed critical _ ` ageing ' _ ( for more details , see refs .
@xcite ) . the field - theoretic treatment of the model a / b dynamical action ( [ i2hajd ] ) , ( [ i2ndjd ] ) with the initial term ( [ i7init ] ) follows the theory of _ boundary critical phenomena _
however , it turns out that additional singularities on the temporal ` surface ' at @xmath390 appear only for model a , and can be incorporated into a single new renormalisation factor ; to one - loop order , one finds @xcite @xmath391 this in turn leads to a _
single _ independent critical exponent associated with the initial time relaxation , the _ initial slip exponent _ , which becomes for the purely relaxational models a and b with nonconserved and conserved order parameter : @xmath392 in order to obtain the _ short - time _ scaling laws for the dynamic response and correlation functions in the _ ageing limit _
@xmath393 , one requires additional information that can be garnered from the short - distance _ operator product expansion _ for the fields , @xmath394 subsequent analysis then yields eventually @xcite @xmath395 and for the time dependence of the mean order parameter @xmath396 one may also compute the universal _ fluctuation dissipation ratios _ in this nonequilibrium ageing regime @xcite .
it emerges , though , that these depend on the quantity under investigation , which prohibits a unique definition of an effective nonequilibrium temperature for critical ageing .
the method sketched above can be extended to models with reversible mode - couplings @xcite . for model j capturing the critical dynamics of isotropic ferromagnets
, one finds @xmath397 in systems where a _
nonconserved _ order parameter is dynamically coupled to other conserved modes , the initial slip exponent @xmath398 is actually _ not _ a universal number , but depends on the width of the initial distribution @xcite .
next we address the question @xcite , what happens if the detailed balance conditions ( [ i1eins ] ) and ( [ i6divf ] ) are violated ?
to start , we change the noise strength @xmath399 in the purely relaxational models a and b , which ( in our units ) violates the einstein relation ( [ i1eins ] ) .
however , this modification can obviously be absorbed into a rescaled _ effective temperature _
, @xmath400 .
formally this is established by means of the dynamical action ( [ i2jade ] ) , which now reads @xmath401 & = & \int \ ! \d^dx \ ! \int \ !
\d t \sum_\alpha { \widetilde s}^\alpha \biggl [ \partial_t \ , s^\alpha + d \ , ( i \vec{\nabla})^a \left ( r - \vec{\nabla}^2 \right ) s^\alpha \nonumber \\ & & \qquad\qquad - { \widetilde d } \ , ( i \vec{\nabla})^a \ , { \widetilde s}^\alpha + d \ , \frac{u}{6 } \ , ( i \vec{\nabla})^a \ , s^\alpha \sum_\beta s^\beta s^\beta \biggr ] \ . \label{i8neab}\end{aligned}\ ] ] upon simple rescaling @xmath402 , @xmath403 , the response functional ( [ i8neab ] ) recovers its equilibrium form , albeit with modified nonlinear coupling @xmath404 . however , the _
universal asymptotic _ properties of these models are governed by the heisenberg fixed point ( [ i5ufix ] ) , and the specific value of the ( renormalised ) coupling , which only serves as the initial condition for the rg flow , does not matter .
in fact , the relaxational dynamics of the kinetic ising model with glauber dynamics ( model a with @xmath213 ) is known to be quite stable against nonequilibrium perturbations @xcite , even if these break the ising @xmath405 symmetry @xcite . for model j the above rescaling modifies in a similar manner merely the mode - coupling strength in eq .
( [ i6mcpl ] ) , namely @xmath406 @xcite .
again , since the dynamic critical behaviour is governed by the universal fixed point ( [ i6mcfp ] ) , thermal equilibrium becomes effectively _ restored _ at criticality .
more generally , it has been established that _
isotropic _ detailed balance violations do not affect the universal properties in other models for critical dynamics that contain additional conserved variables either : the equilibrium rg fixed points tend to be asymptotically stable @xcite . in systems with _
conserved _ order parameter , however , we may in addition introduce spatially _ anisotropic _ violations of einstein s relation ; for example , in model b one can allow for anisotropic relaxation @xmath407 , with different rates in two spatial subsectors and concomitantly anisotropic noise correlations @xmath408 .
we have thus produced a truly nonequilibrium situation provided @xmath409 , which we may interpret as having effectively coupled the longitudinal and transverse spatial sectors to heat baths with different temperatures @xmath410 , say @xcite .
evaluating the fluctuation - induced shift of the transition temperature , see eq .
( [ i4cpsh ] ) one finds not surprisingly that the _ transverse _ sector softens first , while the longitudinal sector remains noncritical .
this suggests that we can neglect the nonlinear longitudinal fluctuations as well as the @xmath411 term in the propagator .
these features are indeed encoded in the corresponding _ anisotropic _ scaling : @xmath412 = \mu$ ] , @xmath413 = \mu^2 $ ] , @xmath224 = \mu^4 $ ] , whence @xmath414 = [ d_\perp ] = \mu^0 $ ] , and @xmath415 = [ d_\parallel ] = \mu^{-2}$ ] become irrelevant . upon renaming @xmath416 and @xmath417
, this ultimately leads to the _ randomly driven _ or _ two - temperature model b _ @xcite as the effective theory describing the phase transition : @xmath418 s^\alpha(\vec{x},t ) \nonumber \\ & & + \frac{d \ , \widetilde u}{6 } \ ,
\vec{\nabla}_\perp^2 \ , s^\alpha(\vec{x},t ) \sum_\beta [ s^\beta(\vec{x},t)]^2 + \zeta^\alpha(\vec{x},t ) \ , \label{i8ttmb}\end{aligned}\ ] ] with the noise correlations @xmath419 quite remarkably , the langevin equation ( [ i8ttmb ] ) can be derived as an _ equilibrium _ diffusive relaxational kinetics @xmath420}{\delta s^\alpha(\vec{x},t ) } + \zeta^\alpha(\vec{x},t ) \label{i8ttrd}\ ] ] from an effective _ long - range _
hamiltonian @xmath421 & = & \int \frac{\d^dq}{(2 \pi)^d } \frac{\vec{q}_\perp^2 ( r + \vec{q}_\perp^2 ) + c \ , \vec{q}_\parallel^2}{2 \ , \vec{q}_\perp^2 } \ , \sum_\alpha \big| s^\alpha(\vec{q } ) \big|^2 \nonumber \\ & & + \frac{\widetilde u}{4 ! } \int \ !
\d^dx \sum_{\alpha,\beta }
[ s^\alpha(\vec{x})]^2 \ , [ s^\beta(\vec{x})]^2 \ .
\label{i8effh}\end{aligned}\ ] ] power counting gives @xmath422 = \mu^{4 - d_\parallel - d}$ ] : the spatial anisotropy suppresses longitudinal fluctuations and lower the upper critical dimension to @xmath423 .
the anisotropic correlations encoded in eq .
( [ i8effh ] ) also reduce the lower critical dimension and affect the nature of the ordered phase @xcite .
the scaling law for , e.g. , the dynamic response function takes the form @xmath424 where we have introduced a new _ anisotropy exponent _ @xmath384 .
since the nonlinear coupling @xmath282 only affects the transverse sector , we find to _ all _ orders in the perturbation expansion : @xmath425 and consequently obtain the @xmath243 factor identity @xmath426 which at any ir - stable rg fixed point implies the _ exact _ scaling relations @xmath427 whereas the scaling exponents for the longitudinal sector read @xmath428 as for the equilibrium model b , the only independent critical exponents to be determined are @xmath14 and @xmath9 . to one - loop order ,
only the combinatorics of the feynman diagrams ( see fig . [ temver ] ) enters their explicit values , whence one finds for @xmath429 formally identical results as for the usual ginzburg
wilson hamiltonian ( [ i1glwh ] ) , @xmath430 albeit with _ different _ @xmath431 . to two - loop order
, however , the anisotropy manifestly affects the evaluation of the loop contributions , and the value for @xmath14 deviates from the expression in eq .
( [ i5stex ] ) @xcite .
interestingly , an analogously constructed nonequilibrium _ two - temperature model j _ with reversible mode - coupling vertex _ can not _ be cast into a form that is equivalent to an equilibrium system , for owing to the emerging anisotropy , the condition ( [ i6divf ] ) can not be satisfied
. a one - loop rg analysis yields a runaway flow , and no stable rg fixed point is found @xcite .
similar behaviour ensues in other anisotropic nonequilibrium variants of critical dynamics models with conserved order parameter ; the precise interpretation of the apparent instability is as yet unclear @xcite . finally , we wish to consider langevin representations of genuinely nonequilibrium systems , namely driven diffusive lattice gases ( for a comprehensive overview , see ref .
first we address the coarse - grained continuum version of the asymmetric exclusion process , i.e. , hard - core repulsive particles that hop preferentially in one direction .
we describe this system in terms of a _ conserved _ particle density , whose fluctuations we denote with @xmath432 , such that @xmath6 , obeying a _ continuity equation _ @xmath433 .
we assume the system to be driven along the ` @xmath434 ' direction ; in the transverse sector ( of dimension @xmath435 ) we thus just have a noisy _ diffusion current _ @xmath436 , whereas there is a nonlinear term , stemming from the hard - core interactions , in the current along the direction of the external drive , with @xmath437 : @xmath438 .
for the stochastic currents , we assume gaussian white noise @xmath439 and @xmath440 , @xmath441 .
notice that since we are not in thermal equilibrium , einstein s relation need not be fulfilled .
we can however always rescale the field to satisfy it in the transverse sector ; the ratio @xmath442 then measures the deviation from equilibrium .
these considerations yield the generic langevin equation for the density fluctuations in _ driven diffusive systems ( dds ) _
@xcite @xmath443 with conserved noise @xmath444 , where @xmath445 and @xmath446 notice that the drive term @xmath447 breaks both the system s spatial reflection symmetry and the ising @xmath405 symmetry @xmath448 .
the corresponding janssen de dominicis response functional ( [ i2jade ] ) reads @xmath401 & = & \int \
! \int \ !
\d t \ , { \widetilde s } \biggl [ \frac{\partial s}{\partial t } - d \left ( \vec{\nabla}_\perp^2 + c \ , \nabla_\parallel^2 \right ) s \nonumber \\ & & \qquad\qquad\qquad + d \left ( \vec{\nabla}_\perp^2 + { \tilde c } \
, \nabla_\parallel^2 \right ) { \widetilde s } - \frac{d \ , g}{2 } \ , \nabla_\parallel \ , s^2 \biggr ] \ . \label{i9ddsf}\end{aligned}\ ] ] it describes a _ ` massless ' _ theory , hence we expect the system to be _ generically scale - invariant _ , without the need to tune it to a special point in parameter space
. the nonlinear drive term will induce anomalous scaling in the drive direction , different from ordinary diffusive behaviour . in the transverse sector , however , we have to _ all _ orders in the perturbation expansion simply @xmath449 since the nonlinear three - point vertex , which is of the form depicted in fig . [ modelj](a ) , is proportional to @xmath450 .
consequently , @xmath451 which immediately implies @xmath452 moreover , the nonlinear coupling @xmath347 itself does not renormalise either as a consequence of _
galilean invariance_. namely , the langevin equation ( [ i9ddsl ] ) and the action ( [ i9ddsf ] ) are left invariant under galilean transformations @xmath453 thus , the boost velocity @xmath454 must scale as the field @xmath455 under renormalisation , and since the product @xmath456 must be invariant under the rg , this leaves us with @xmath457 the effective nonlinear coupling governing the perturbation expansion in terms of loop diagrams turns out to be @xmath458 ; if we define its renormalised counterpart as @xmath459 with the convenient choice @xmath460 , we see that the associated rg beta function becomes @xmath461 at _ any _ nontrivial rg fixed point @xmath462 , therefore @xmath463 .
we thus infer that below the upper critical dimension @xmath464 for dds , the longitudinal scaling exponents are fixed by the system s symmetry @xcite , @xmath465 an explicit one - loop calculation for the two - point vertex functions , see fig .
[ modelj](b ) and ( c ) , yields @xmath466 this establishes that in fact the fixed point @xmath467 is ir - stable ( provided @xmath462 ) , which means that asymptotically the einstein relation is satisfied in the longitudinal sector as well @xcite . in this context
, it is instructive to make an intriguing connection with the _ noisy burgers equation _ @xcite , describing simplified fluid dynamics in terms of a velocity field @xmath468 : @xmath469 = d \ , \vec{\nabla}^2 \vec{u}(\vec{x},t ) + \vec{\zeta}(\vec{x},t ) \ , \label{i9nobg } \\ & & \left\langle \zeta_i \right\rangle = 0 \ , , \
\left\langle \zeta_i(\vec{x},t ) \ , \zeta_j(\vec{x}',t ' ) \right\rangle = - 2 d \ , \vec{\nabla}_i \vec{\nabla}_j \ , \delta(\vec{x}-\vec{x } ' ) \ , \delta(t - t')\ . \label{i9bgno}\end{aligned}\ ] ] for @xmath470 , the nonlinearity is just the usual fluid advection term . in one dimension ,
the burgers equation ( [ i9nobg ] ) becomes _ identical _ with the dds langevin equation ( [ i9ddsl ] ) , so we immediately infer its anomalous dynamic critical exponent @xmath471 . at least in one dimension
therefore , it should represent an _ equilibrium _ system which asymptotically approaches the canonical distribution ( [ i1eqpd ] ) , where the hamiltonian is simply the fluid s kinetic energy ( and we have set @xmath151 ) .
so let us check the equilibrium condition ( [ i6divf ] ) with @xmath472 \propto \exp \left [ - \frac12 \int \ !
\vec{u}(\vec{x})^2 \ , \d^dx \right]$ ] : @xmath473 \e^{- \frac12 \int \ !
\vec{u}(\vec{x}',t)^2 \ , \d^dx ' } \\ & & \quad = \int \ ! \left [ 2 \vec{\nabla } \cdot \vec{u}(\vec{x},t )
- \vec{u}(\vec{x},t ) \cdot \vec{\nabla } \vec{u}(\vec{x},t)^2 \right ] \d^dx \ , \e^{- \frac12 \int \ ! \vec{u}(\vec{x}',t)^2 \ , \d^dx ' } \ .
\end{aligned}\ ] ] with appropriate boundary conditions , the first term here vanishes , but the second one does so _ only _ in @xmath474 : @xmath475
. driven diffusive systems in one dimension are therefore subject to a _ ` hidden ' fluctuation
dissipation theorem_. to conclude this part on langevin dynamics , let us briefly consider the _ driven model b _ or _
critical dds _
@xcite , which corresponds to a driven ising lattice gas near its critical point . here
, a conserved scalar field @xmath455 undergoes a second - order phase transition , but similar to the randomly driven case , again only the _ transverse _ sector is critical . upon adding the dds drive term from eq .
( [ i9ddsl ] ) to the langevin equation ( [ i8ttmb ] ) , we obtain @xmath476 s(\vec{x},t ) + \frac{d \ , \widetilde u}{6 } \ , \vec{\nabla}_\perp^2 \
, s(\vec{x},t)^3 \nonumber \\ & & + \frac{d \ , g}{2 } \ , \nabla_\parallel \ , s(\vec{x},t)^2 + \zeta(\vec{x},t ) \ , \label{i9drmb}\end{aligned}\ ] ] with the ( scalar ) noise specified in eq .
( [ i8ttnc ] ) .
the response functional thus becomes @xmath401 & = & \int \
\d t \ , { \widetilde s } \biggl [ \frac{\partial s}{\partial t } - d \left [ \vec{\nabla}_\perp^2 \left ( r - \vec{\nabla}_\perp^2 \right ) + c \ , \nabla_\parallel^2\right ] s \nonumber \\ & & \qquad\qquad\qquad + d \left ( \vec{\nabla}_\perp^2 \ , { \widetilde s } - \frac{\widetilde u}{6 } \ , \vec{\nabla}_\perp^2 \ , s^3 - \frac{g}{2 } \ , \nabla_\parallel \ , s^2 \right ) \biggr ] \ .
\label{i9dmbf}\end{aligned}\ ] ] power counting gives @xmath477 = \mu^{5-d}$ ] , so the upper critical dimension here is @xmath478 , and @xmath422 = \mu^{3-d}$ ] .
the nonlinearity @xmath479 is thus _ irrelevant _ and can be omitted if we wish to determine the asymptotic universal scaling laws ; but recall that it is responsible for the phase transition in the system .
the remaining vertex is then proportional to @xmath450 , whence eqs .
( [ i9ga11 ] ) and ( [ i9zfid ] ) hold for critical dds as well , and the transverse critical exponents are just those of the gaussian model b , @xmath480 in addition , galilean invariance with respect to eq .
( [ i9galt ] ) and therefore eq . ( [ i9ginv ] )
hold as before . with the renormalised nonlinear drive strength defined similarly to eq .
( [ i9renc ] ) , but a different geometric constant and the scale factor @xmath481 , the associated rg beta function reads @xmath482 which again allows us to determine the longitudinal scaling exponents to _ all _ orders in perturbation theory , for @xmath483 , @xmath484 it is worthwhile mentioning a few marked differences to the two - temperature model b discussed in sec .
[ ssec : i8 ] : in dds , there are obviously nonzero _ three - point _
correlations , and in the driven critical model b the upper critical dimension is @xmath478 as opposed to @xmath423 for the randomly driven version .
notice also that the latter is characterised by nontrivial static critical exponents , but the kinetics is purely diffusive along the drive direction , @xmath485 .
conversely for the driven model b , only the longitudinal scaling exponents are non - gaussian .
we now turn our attention to stochastic interacting particle systems , whose microscopic dynamics is defined through a ( classical ) master equation .
below , we shall see how the latter can be mapped onto a stochastic quasi - hamiltonian in a second - quantised bosonic operator representation @xcite .
taking the continuum on the basis of coherent - state path integrals then yields a field theory action that may be analysed by the very same rg methods as described before in secs .
[ ssec : i3][ssec : i5 ] ( for more details , see the recent overview @xcite ) .
our goal is to study systems of ` particles ' @xmath486 that propagate through hopping to nearest neighbors on a @xmath13-dimensional lattice , or via diffusion in the continuum . upon encounter , or spontaneously , with given stochastic rates , these particles may undergo species changes , annihilate , or produce offspring . at large densities ,
the characteristic time scales of the kinetics will be governed by the reaction rates , and the system is said to be _ reaction - limited_. in contrast , at low densities , any reactions that require at least two particles to be in proximity will be _ diffusion - limited _ : the basic time scale will be set by the hopping rate or diffusion coefficient . as a first approximation to the dynamics of such ` chemical ' reactions , let us assume homogeneous mixing of each species .
we may then hope to be able to capture the kinetics in terms of _ rate equations _ for each particle concentration or mean density .
note that such a description neglects any spatial fluctuations and correlations in the system , and is therefore in character a mean - field approximation . as a first illustration
consider the _ annihilation _ of @xmath487 particles of species @xmath488 in the _ irreversible _ @xmath186th - order reaction @xmath489 , with rate @xmath490 .
the corresponding rate equation employs a factorisation of the probability of encountering @xmath186 particles at the same point to simply the @xmath186th power of the concentration @xmath491 , @xmath492 this ordinary differential equation is readily solved , with the result @xmath493^{-1 /(k-1 ) } \ .
\label{ii1mfpd}\end{aligned}\ ] ] for simple ` radioactive ' decay ( @xmath494 ) , we of course obtain an exponential time dependence , as appropriate for statistically independent events . for pair ( @xmath495 ) and
higher - order ( @xmath496 ) processes , however , we find algebraic long - time behaviour , @xmath497 , with an amplitude that becomes independent of the initial density @xmath498 .
the absence of a characteristic time scale hints at cooperative effects , and we have to ask if and under which circumstances correlations might qualitatively affect the asymptotic long - time power laws . for according to smoluchowski theory @xcite , we would expect the annihilation reactions to produce _ depletion zones _ in sufficiently low dimensions @xmath1 , which would in turn induce a considerable _ slowing down _ of the density decay , see sec .
[ ssec : ii3 ] . for _ two - species _ pair annihilation @xmath499 ( without mixing ) , another complication emerges , namely particle species _ segregation _ in dimensions for @xmath500 ; the regions dominated by either species become largely inert , and the annihilation reactions are confined to rather sharp _
fronts _ @xcite .
_ competition _ between particle decay and production processes , e.g. , in the reactions @xmath501 ( with rate @xmath502 ) , @xmath503 ( with forward and back rates @xmath228 and @xmath490 , respectively ) , leads to even richer scenarios , as can already be inferred from the associated rate equation @xmath504 for @xmath505 , clearly @xmath506 as @xmath336 .
the system eventually enters an _ inactive _ state , which even in the fully stochastic model is _ absorbing _ , since once there is no particle left , no process whatsoever can drive the system out of the empty state again . on the other hand , for @xmath507
, we encounter an _ active _ state with @xmath508 exponentially , with rate @xmath509 .
we have thus identified a nonequilibrium _ continuous _ phase transition at @xmath510 . indeed , as in equilibrium critical phenomena , the critical point is governed by characteristic power laws ; for example , the asymptotic particle density @xmath511 , and the critical density decay @xmath512 with @xmath513 in the mean - field approximation .
the following natural questions then arise : what are the _ critical exponents _ once statistical fluctuations are properly included in the analysis ?
can we , as in equilibrium systems , identify and characterise certain _ universality classes _ , and which microscopic or overall , global features determine them and their critical dimension ?
already the previous set of reactions may also be viewed as a ( crude ) model for the _ population dynamics _ of a single species . in the same language
, we may also formulate a stochastic version of the classic _ lotka volterra predator prey competition model _
@xcite : if by themselves , the ` predators ' @xmath488 die out according to @xmath501 , with rate @xmath502 , whereas the prey reproduce @xmath514 with rate @xmath228 , and thus proliferate with a malthusian population explosion .
the predators are kept alive and the prey under control through _ predation _ , here modeled as the reaction @xmath515 : with rate @xmath490 , a prey is ` eaten ' by a predator , who simultaneously produces an offspring .
the coupled kinetic rate equations for this system read @xmath516 it is straightforward to show that the quantity @xmath517 - \sigma \ln{a(t ) } - \kappa \ln{b(t)}$ ] is a constant of motion for this coupled system of differential equations , i.e. , @xmath518 . as a consequence , the system is governed by regular population _
oscillations _ , whose frequency and amplitude are fully determined by the _ initial _ conditions . clearly , this is not a very realistic feature ( albeit mathematically appealing ) , and moreover eqs .
( [ ii1lvre ] ) are known to be quite unstable with respect to model modifications @xcite .
indeed , if one includes spatial degrees of freedom and takes account of the full stochasticity of the processes involved , the system s behaviour turns out to be much richer @xcite : in the species coexistence phase , one encounters for sufficiently large values of the predation rate an incessant sequence of _ ` pursuit and evasion ' _ waves that form quite complex dynamical patterns .
in finite systems , these induce _ erratic _ population oscillations whose features are however independent of the initial configuration , but whose amplitude vanishes in the thermodynamic limit .
moreover , if locally the prey ` carrying capacity ' is limited ( corresponding to restricting the maximum site occupation number per site on a lattice ) , there appears an _ extinction threshold _ for the predator population that separates the absorbing state of a system filled with prey from the active coexistence regime through a continuous phase transition @xcite .
these examples all call for a systematic approach to include stochastic fluctuations in the mathematical description of interacting reaction diffusion systems that would be conducive to the application of field - theoretic tools , and thus allow us to bring the powerful machinery of the dynamic renormalisation group to bear on these problems . in the following
, we shall describe such a general method @xcite which allows a representation of the classical master equation in terms of a coherent - state path integral and its subsequent analysis by means of the rg ( for overviews , see refs .
@xcite ) .
the above interacting particle systems , when defined on a @xmath13-dimensional lattice with sites @xmath519 , are fully characterised by the set of occupation integer numbers @xmath520 for each particle species .
the _ master equation _ then describes the temporal evolution of the configurational probability distribution @xmath521 through a _ balance _ of gain and loss terms .
for example , for the _ binary annihilation _ and _ coagulation reactions _
@xmath522 with rate @xmath490 and @xmath523 with rate @xmath524 , the master equation on a specific site @xmath519 reads @xmath525 with initially @xmath526 , e.g. , a _ poisson _ distribution @xmath527 .
since the reactions all change the site occupation numbers by integer values , a _
second - quantised fock space representation _ is particularly useful @xcite . to this end
, we introduce the _ bosonic operator algebra _ @xmath528 = 0 = \bigl [ a_i^\dagger , a_j^\dagger \bigr ] \ , \quad \bigl [ a_i , a_j^\dagger \bigr ] = \delta_{ij } \ . \label{ii2boso}\ ] ] from these commutation relations one establishes in the standard manner that @xmath529 and @xmath530 constitute lowering and raising ladder operators , from which we may construct the particle number eigenstates @xmath531 , @xmath532 ( notice that we have chosen a different normalisation than in ordinary quantum mechanics . ) a state with @xmath533 particles on sites @xmath519 is then obtained from the empty vaccum state @xmath534 , defined through @xmath535 , as the product state @xmath536 to make contact with the time - dependent configuration probability , we introduce the formal _ state vector _ @xmath537 whereupon the linear time evolution according to the master equation is translated into an _ ` imaginary - time ' schrdinger equation _ @xmath538 the _ stochastic quasi - hamiltonian _ ( rather , the time evolution or liouville operator ) for the on - site reaction processes is a sum of local terms , @xmath539 ; e.g. , for the binary annihilation and coagulation reactions , @xmath540 the two contributions for each process may be physically interpreted as follows : the first term corresponds to the actual _ process _ under consideration , and describes how many particles are annihilated and ( re-)created in each reaction .
the second term gives the _ ` order ' _ of each reaction , i.e. , the number operator @xmath541 appears to the @xmath186th power , but in normal - ordered form as @xmath542 , for a @xmath186th - order process .
note that the reaction hamiltonians such as ( [ ii2hrea ] ) are _ non - hermitean _ , reflecting the particle creations and destructions . in a similar manner ,
hopping between neighbouring sites @xmath543 is represented in this formalism through @xmath544 our goal is of course to compute averages with respect to the configurational probability distribution @xmath521 ; this is achieved by means of the _ projection state _ @xmath545 , which satisfies @xmath546 and @xmath547 , since @xmath548 = \e^{a_i } \ , \delta_{ij}$ ] . for the desired _ statistical averages _ of observables that must be expressible in terms of the occupation numbers @xmath549 ,
we then obtain @xmath550 let first us explore the consequences of _ probability conservation _
, i.e. , @xmath551 .
this requires @xmath552 ; upon commuting @xmath553 with @xmath554 , effectively the creation operators become shifted @xmath555 , whence this condition is fulfilled provided @xmath556 , which is indeed satisfied by our explicit expressions ( [ ii2hrea ] ) and ( [ ii2hdif ] ) . by this prescription ,
we may also in averages replace @xmath557 , i.e. , the _ particle density _ becomes @xmath558 , and the two - point operator @xmath559 . in the bosonic operator representation
above , we have assumed that there exist no restrictions on the particle occupation numbers @xmath533 on each site .
if , however , there is a maximum @xmath560 , one may instead employ a representation in terms of spin @xmath561 operators .
for example , particle exclusion systems with @xmath562 or @xmath563 can thus be mapped onto non - hermitean spin @xmath564 ` quantum ' systems .
specifically in one dimension , such representations in terms of integrable spin chains have proved a fruitful tool ; for overviews , see refs .
an alternative approach uses the bosonic theory , but encodes the site occupation restrictions through appropriate exponentials in the number operators @xmath565 @xcite .
we may now follow an established route in quantum many - particle theory @xcite and proceed towards a field theory representation through constructing the _ path integral _ equivalent to the ` schrdinger ' dynamics ( [ ii2schr ] ) based on _ coherent states _
, which are right eigenstates of the annihilation operator , @xmath566 , with complex eigenvalues @xmath567 .
explicitly , one finds @xmath568 satisfying the _ overlap _ and ( over-)_completeness relations _
@xmath569 upon splitting the temporal evolution ( [ ii2schr ] ) into infinitesimal steps , and inserting eqs .
( [ ii2ovcp ] ) at each time step , standard procedures ( elaborated in detail in ref .
@xcite ) yield eventually @xmath570 \
, { \cal d}[\phi_i^ * ] \ , f(\ { \phi_i \ } ) \ , \e^{- { \cal a}[\phi_i^*,\phi_i ] } \ , \label{ii2copi}\ ] ] with the _ action _ @xmath571 = \sum_i \biggl ( - \phi_i(t_f ) + \int_0^{t_f } \ ! \d t \left [ \phi_i^ * \ , \frac{\partial \phi_i}{\partial t } + h_i(\phi_i^*,\phi_i ) \right ] - { \bar n}_0 \ , \phi^*_i(0 )
\biggr ) \ , \label{ii2cact}\ ] ] where the first term originates from the projection state , and the last one from the initial poisson distribution .
notice that in the hamiltonian , the creation and annihilation operators @xmath530 and @xmath529 are simply replaced with the complex numbers @xmath572 and @xmath567 , respectively . taking the _ continuum limit _
, @xmath573 , @xmath574 , the ` bulk ' part of the action becomes @xmath575 = \int \ ! \d^dx \ ! \int \ ! \d t \left [ { \hat \psi } \left ( \frac{\partial}{\partial t } - d \ , \vec{\nabla}^2 \right ) \psi + { \cal h}_{\rm reac}({\hat \psi},\psi ) \right ] \ , \label{ii2maft}\ ] ] where the hopping term ( [ ii2hdif ] ) has naturally turned into a diffusion propagator .
we have thus arrived at a _ microscopic _ stochastic field theory for reaction
diffusion processes , with _ no _ assumptions whatsoever on the form of the ( internal ) noise .
this is a crucial ingredient for nonequilibrium dynamics , and we may now use eq .
( [ ii2maft ] ) as a basis for systematic coarse - graining and the renormalisation group analysis . returning to our example of
pair annihilation and coagulation , the reaction part of the action ( [ ii2maft ] ) reads @xmath576 see eq .
( [ ii2hrea ] ) .
let us have a look at the _ classical field equations _ , namely @xmath577 , which is always solved by @xmath578 , reflecting probability conservation , and @xmath579 , which , upon inserting @xmath578 gives here @xmath580 i.e. , essentially the mean - field rate equation for the local particle density @xmath581 , see eq .
( [ ii1mfan ] ) , supplemented with diffusion .
the field theory action ( [ ii2maft ] ) , derived from the master equation ( [ ii2mast ] ) , then provides a means of including fluctuations in our analysis . before we proceed with this program , it is instructive to perform a shift in the field @xmath582 about the mean - field solution , @xmath583 , whereupon the reaction hamiltonian density ( [ ii2anco ] ) becomes @xmath584 in addition to the diffusion propagator , the annihilation and coagulation processes thus give _ identical _ three- and four - point vertices ; aside from non - universal amplitudes , one should therefore obtain _ identical _ scaling behaviour for both binary reactions in the asymptotic long - time limit @xcite .
lastly , we remark that if we interpret the action @xmath585 $ ] as a response functional ( [ i2jade ] ) , despite the fields @xmath586 not being purely imaginary , our field theory becomes formally equivalent to a ` langevin ' equation , wherein additive noise is added to eq .
( [ ii2dreq ] ) , albeit with _ negative _ correlator @xmath587 = - ( \lambda + \lambda ' ) \ , \psi^2 $ ] , which represents _ ` imaginary ' multiplicative noise_. this langevin description is thus not well - defined ; however , one may render the noise correlator positive through a nonlinear _ cole hopf transformation _
@xmath588 , @xmath589 such that @xmath590 , with jacobian @xmath563 , but at the expense of ` diffusion noise ' @xmath591 in the action @xcite . in summary , binary ( and higher - order ) annihilation and coagulation processes can not be cast into a langevin framework in any simple manner .
we begin by analysing diffusion - limited single - species annihilation @xmath592 @xcite .
the corresponding field theory action ( [ ii2maft ] ) reads @xmath575 = \int \ ! \d^dx \
\d t \left [ { \hat \psi } \left ( \frac{\partial}{\partial t } - d \vec{\nabla}^2 \right ) \psi - \lambda \left ( 1 - { \hat \psi}^k \right ) \psi^k \right ] \ , \label{ii3anac}\ ] ] which for @xmath496 allows no ( obvious ) equivalent langevin description .
straightforward power counting gives the scaling dimension for the annihilation rate , @xmath593 = \mu^{2 - ( k-1 ) d}$ ] , which suggests the upper critical dimension @xmath594 .
thus we expect mean - field behaviour @xmath595 , see eq .
( [ ii1mfpd ] ) , in any physical dimension for @xmath596 , logarithmic corrections at @xmath597 for @xmath598 and at @xmath464 for @xmath495 , and nonclassical power laws for pair annihilation only in one dimension .
the field theory defined by the action ( [ ii3anac ] ) has two vertices , the ` annihilation ' sink with @xmath186 incoming lines only , and the ` scattering ' vertex with @xmath186 incoming and @xmath186 outgoing lines .
neither allows for propagator renormalisation , hence the model remains massless with _
exact _ scaling exponents @xmath379 and @xmath599 , i.e. , diffusive dynamics .
in addition , the entire perturbation expansion for the renormalisation for the annihilation vertices is merely a geometric series of the one - loop diagram , see fig .
[ versum ] for the pair annihilation case ( @xmath495 ) .
if we define the renormalised effective coupling according to @xmath600 where @xmath601 , we obtain for the single nontrivial renormalisation constant @xmath602 to _ all _ orders . consequently , the associated rg beta function becomes @xmath603 with the gaussian fixed point @xmath604 stable for @xmath605 , leading to the mean - field power laws ( [ ii1mfpd ] ) , whereas for @xmath606 the flow approaches @xmath607 since the particle density has scaling dimension @xmath608 = \mu^d$ ] , we may write @xmath609 , where we have retained the dependence on the initial density @xmath610 .
since the fields and the diffusion constant do not renormalise ( @xmath611 and @xmath612 ) , the rg equation for the density takes the form @xmath613 { \hat a}_r\left ( n_0 \ , \mu^{-d } , g_r \right ) = 0 \ , \label{ii3rgeq}\ ] ] see eq .
( [ i5rgsu ] ) . with
the characteristics set equal to @xmath614 , the solution of the rg equation ( [ ii3rgeq ] ) near the ir - stable rg fixed point @xmath615 becomes @xmath616 under the rg , the first argument in eq .
( [ ii3drgs ] ) flows to infinity .
one therefore needs to establish that the result for the scaling function @xmath617 is finite to _ all _ orders in the initial density @xcite .
one then finds the following asymptotic long - time behaviour for pair annihilation below the critical dimension @xcite , @xmath618 at the critical dimension , @xmath619 logarithmically slowly , and the process is still diffusion - limited ; this gives @xmath620^{1/2 } \ .
\label{ii3tdlc } \end{aligned}\ ] ] in pair annihilation reactions of two _ distinct _
species @xmath499 , where _ no _ reactions between the same species are allowed , a novel phenomenon emerges in sufficiently low dimensions @xmath500 , namely particle segregation in separate spatial domains , with the decay processes restricted to sharp reaction fronts on their boundaries @xcite . note that the reaction @xmath499 preserves the difference of particle numbers ( even locally ) , i.e. , there is a _ conservation law _ for @xmath621 @xcite .
the rate equations for the concentrations @xmath622 are for _ equal _ initial densities @xmath623 solved by the single - species pair annihilation mean - field power law @xmath624 whereas for _ unequal _ initial densities @xmath625 , say , the majority species @xmath626 as @xmath336 , and the minority population disappears , @xmath627 . from eqs . ( [ ii4mfre ] )
we obtain for @xmath628 the exponential approach @xmath629 mapping the associated master equation onto a continuum field theory ( [ ii2maft ] ) , the reaction term now reads ( with the fields @xmath630 and @xmath631 representing the @xmath488 and @xmath632 particles , respectively ) @xcite @xmath633 as in the single - species case , there is no propagator renormalisation , and moreover the feynman diagrams for the renormalised reaction vertex are of precisely the same form as for @xmath522 , see fig .
[ versum ] .
thus , for unequal initial densities , @xmath634 , the mean - field power law @xmath635 in the exponent of eq .
( [ ii4udms ] ) becomes again replaced with @xmath636 in dimensions @xmath637 , leading to _ stretched exponential _ time dependence , @xmath638 however , species segregation for _ equal _ initial densities , @xmath623 , even supersedes the slowing down due to the reaction rate renormalisation . as confirmed by a thorough rg analysis , this effect can be captured within the classical field equations @xcite . to this end , we add _ diffusion _ terms ( with equal diffusivities ) to the rate equations ( [ ii4mfre ] ) for the now _ local _ particle densities , @xmath639 the local concentration difference @xmath640 thus becomes a purely _ diffusive mode _
, @xmath641 , and we employ the _ diffusion green function _ @xmath642 compare eq .
( [ i1rest ] ) , to solve the initial value problem , @xmath643 let us furthermore assume a _
poisson distribution _ for the _ initial _ density correlations ( indicated by an overbar ) , @xmath644 and @xmath645 , which implies @xmath646 .
averaging over the initial conditions then yields with eq .
( [ ii4iniv ] ) @xmath647 since the distribution for @xmath648 will be a gaussian , we thus obtain for the _ local density excess _ originating in a random initial fluctuation , @xmath649 in dimensions @xmath650 these density fluctuations decay _ slower _ than the overall particle number @xmath651 for @xmath652 and @xmath653 for @xmath654 in a homogeneous system .
species segregation into @xmath488- and @xmath632-rich domains renders the particle distribution _ nonuniform _ , and the density decay is governed by the slow power law ( [ ii4segd ] ) , @xmath655 .
for very _ special _ initial states , however , the situation can be different .
for example , consider hard - core particles ( or @xmath656 ) regularly arranged in an alternating manner @xmath657 on a one - dimensional chain .
the reactions @xmath499 preserve this arrangement , whence the distinction between @xmath488 and @xmath632 particles becomes meaningless , and one indeed recovers the @xmath658 power law from the single - species pair annihilation reaction .
let us at last generalise to _
@xmath659-species _ annihilation @xmath660 , with @xmath661 , with equal initial densities @xmath662 as well as uniform diffusion and reaction rates . for @xmath663 ,
there exists _ no _ conservation law in the stochastic system , and one may argue , based on the study of fluctuations in the associated fokker
planck equation , that segregation happens only for @xmath664 @xcite . in any physical dimension @xmath237
, one should therefore see the same behaviour as for the single - species reaction @xmath522 ; this is actually obvious for @xmath665 , since in this case the probability for particles of the same species to ever meet is zero , whence the species labeling becomes irrelevant . in one dimension , with its special topology , segregation does occur , and for generic initial conditions one finds the the decay law @xcite @xmath666 which recovers @xmath667 and @xmath668 . once again , in special situations , e.g. , the alignment @xmath669 for @xmath670 , the single - species scaling ensues .
there are also curious _
variants , for example if for four species we only allow the reactions @xmath499 , @xmath671 , @xmath672 , and @xmath673
. we may then obviously identify @xmath674 and @xmath675 , which leads back to the case of two - species pair annihilation .
generally , within essentially mean - field theory one finds for cyclic multi - species annihilation processes @xmath676 , where for @xmath677 remarkably , for five species this yields the borderline dimension @xmath678 for segregation to occur , hence nontrivial decay exponents @xmath679 in @xmath680 and @xmath681 in @xmath682 that involve the golden ratio @xcite .
let us now return to the _ competing _ single - species reactions @xmath501 ( rate @xmath502 ) , @xmath683 ( rate @xmath228 ) , and , in order to limit the particle density in the active phase , @xmath523 ( rate @xmath490 ) . adding diffusion to the rate equation ( [ ii1mfpt ] ) , we arrive at the _ fisher kolmogorov equation _ of biology and ecology @xcite , @xmath684 where @xmath685 . as discussed in sec .
[ ssec : ii1 ] , it predicts a _ continuous _ transition from an active to an inactive , _ absorbing _ state to occur at @xmath686 .
if we define the associated _ critical exponents _ in close analogy to equilibrium critical phenomena , see sec .
( [ ssec : i1 ] ) , the partial differential equation ( [ ii5fiko ] ) yields the gaussian exponent values @xmath687 , @xmath111 , @xmath113 , and @xmath688 . by the methods outlined in sec .
[ ssec : ii2 ] , we may construct the coherent - state path integral ( [ ii2maft ] ) for the associated master equation , @xmath689 & = & \int \ ! \d^dx \
\biggl [ { \hat \psi } \left ( \frac{\partial}{\partial t } - d \ , \vec{\nabla}^2 \right ) \psi - \kappa \left ( 1 - { \hat \psi } \right ) \psi \nonumber \\ & & \qquad\qquad\quad + \sigma \left ( 1 - { \hat \psi } \right ) { \hat \psi } \ , \psi - \lambda \left ( 1 - { \hat \psi } \right ) { \hat \psi } \ , \psi^2 \biggr ] \ .
\label{ii5mase } \end{aligned}\ ] ] upon shifting the field @xmath582 about its stationary value @xmath563 and rescaling according to @xmath690 and @xmath691 , the action becomes @xmath136 = \int \ ! \d^dx \
\d t \ , \left [ { \widetilde s } \left ( \frac{\partial}{\partial t } + d \left ( r - \vec{\nabla}^2 \right ) \right ) s - u \left ( { \widetilde s } - s \right ) { \widetilde s } \ , s + \lambda \ , { \widetilde s}^2 \ , s^2 \right ] \ . \label{ii5regf}\ ] ] thus , the three - point vertices have been scaled to identical coupling strengths @xmath692 , which represents the effective coupling of the perturbation expansion , see fig . [ dploop ] below .
its scaling dimension is @xmath227 = \mu^{2-d/2}$ ] , whence we infer the upper critical dimension @xmath209 .
the four - point vertex @xmath693 , with @xmath593 = \mu^{2-d}$ ] , is thus _ irrelevant _ in the rg sense , and can be dropped for the computation of universal , asymptotic scaling properties . the action ( [ ii5regf ] ) with @xmath694 is known as _ reggeon field theory _ @xcite , and its basic characteristic is its invariance under _ rapidity inversion _ @xmath695 .
if we interpret eq .
( [ ii5regf ] ) as a response functional , we see that it becomes formally equivalent to a stochastic process with multiplicative noise ( @xmath696 ) captured by the langevin equation @xcite @xmath697 ( for a more accurate mapping procedure , see ref .
@xcite ) , which is essentially a noisy fisher kolmogorov equation ( [ ii5fiko ] ) , with the noise correlator ( [ ii5dpns ] ) ensuring that the fluctuations indeed cease in the absorbing state where @xmath6 .
it has moreover been established @xcite that the action ( [ ii5regf ] ) describes the scaling properties of critical _ directed percolation ( dp ) _
clusters @xcite , illustrated in fig . [ dirpcl ] .
indeed , if the dp growth direction is labeled as ` time ' @xmath698 , we see that the structure of the dp clusters emerges from the basic decay , branching , and coagulation reactions encoded in eq .
( [ ii5mase ] ) .
in fact , the field theory action should govern the scaling properties of _ generic _ continuous nonequilibrium phase transitions from active to inactive , absorbing states , namely for an order parameter with markovian stochastic dynamics that is decoupled from any other slow variable , and in the absence of quenched randomness @xcite .
dp conjecture _ follows from the following _ phenomenological _ approach @xcite to _ simple epidemic processes ( sep ) _ , or epidemics with recovery @xcite : 1 .
a _ ` susceptible ' medium _ becomes locally _ `
infected ' _ , depending on the density @xmath312 of neighboring ` sick ' individuals .
the infected regions recover after a brief time interval .
2 . the state @xmath700 is _
absorbing_. it describes the _ extinction _ of the ` disease ' .
3 . the disease spreads out _ diffusively _ via the short - range infection 1 . of neighboring susceptible regions .
microscopic fast degrees of freedom are incorporated as _ local noise _ or stochastic forces that respect statement 2 .
, i.e. , the noise alone can not regenerate the disease .
these ingredients are captured by the coarse - grained mesoscopic langevin equation @xmath701 \right ) n + \zeta$ ] with a reaction functional @xmath702 $ ] , and s stochastic noise correlator of the form @xmath703 = n \ , n[n]$ ] . near the extinction threshold
, we may expand @xmath702 = r + u \
, n \ , + \ldots$ ] , @xmath704 = v + \ldots$ ] , and higher - order terms turn out to be _ irrelevant _ in the rg sense . upon rescaling
, we recover the reggeon field theory action ( [ ii5regf ] ) for dp as the corresponding response functional ( [ i2jade ] ) .
we now proceed to an explicit evaluation of the dp critical exponents to one - loop order , closely following the recipes given in secs .
[ ssec : i3][ssec : i5 ] .
the lowest - order fluctuation contribution to the two - point vertex function @xmath372 ( propagator self - energy ) is depicted in fig .
[ dploop](a ) .
the feynman rules of sec .
[ ssec : i3 ] yield the corresponding analytic expression @xmath705 the _ criticality condition _
@xmath706 at @xmath707 provides us with the fluctuation - induced shift of the percolation threshold @xmath708 inserting @xmath709 into eq .
( [ ii5ga11 ] ) , we then find to this order @xmath710 and the diagram in fig .
[ dploop](b ) for the three - point vertex functions , evaluated at zero external wavevectors and frequencies , gives @xmath711 for the renormalisation factors , we use again the conventions ( [ i4fren ] ) and ( [ i4pren ] ) , but with @xmath712 because of rapidity inversion invariance , @xmath713 . with eq .
( [ i4dimr ] ) the derivatives of @xmath172 with respect to @xmath714 , @xmath715 , and @xmath71 , as well as the one - loop result for @xmath716 in eq .
( [ ii5gm12 ] ) , all evaluated at the normalisation point @xmath257 , provide us with the @xmath243 factors @xmath717 from these we infer the rg flow functions @xmath718 with the renormalised effective coupling @xmath719 whose rg beta function is to this order @xmath720 \ .
\label{ii5dpbf}\ ] ] for @xmath317 , the gaussian fixed point @xmath721 is stable , and we recover the mean - field critical exponents . for @xmath722 , we find the nontrivial ir - stable rg fixed point @xmath723 setting up and solving the rg equation ( [ i5rgeq ] ) for the vertex function proceeds just as in sec .
[ ssec : i5 ] . with the identifications ( [ i5crex ] ) ( with @xmath92 )
we thus obtain the dp _ critical exponents _ to first order in @xmath724 , @xmath725 in the vicinity of @xmath726 , the solution of the rg equation for the _ order parameter _ reads , recalling that @xmath727 = \mu^{d/2}$ ] , @xmath728 which leads to the following scaling relations and explicit exponent values , @xmath729 llll scaling exponent & @xmath730 & @xmath731 & @xmath732 + @xmath733 & @xmath734 & @xmath735 & @xmath736 + @xmath737 & @xmath738 & @xmath739 & @xmath740 + @xmath741 & @xmath742 & @xmath743 & @xmath744 + @xmath745 & @xmath746 & @xmath747 & @xmath748 + the scaling exponents for critical directed percolation are known analytically for a plethora of physical quantities ( but the reader should beware that various different conventions are used in the literature ) ; for the two - loop results to order @xmath749 in the perturbative dimensional expansion , see ref .
@xcite . in table
[ dpexpt ] , we compare the @xmath750 values with the results from monte carlo computer simulations , which allow the dp critical exponents to be measured to high precision ( for recent overviews on simulation results for dp and other absorbing state phase transitions , see refs .
@xcite ) . yet
unfortunately , there are to date hardly any real experiments that would confirm the dp conjecture @xcite and actually measure the scaling exponents for this prominent nonequilibrium universality class . 1 .
the susceptible medium becomes infected , depending on the densities @xmath312 _ and _ @xmath751 of sick individuals and the _ ` debris ' _ , respectively . after a brief time interval , the sick individuals decay into immune debris , which ultimately stops the disease locally by exhausting the supply of susceptible regions .
the states with @xmath700 and any spatial distribution of @xmath751 are _ absorbing _ , and describe the _ extinction _ of the disease . here
, the debris is given by the accumulated decay products , @xmath752 after rescaling , this _ general epidemic process ( gep ) _ or epidemic with removal @xcite is described in terms of the mesoscopic langevin equation @xcite @xmath753 with noise correlator ( [ ii5dpns ] ) . the associated response functional reads @xcite @xmath136 = \int \
\d t \ , \biggl [ { \widetilde s } \left ( \frac{\partial}{\partial t } + d \left ( r - \vec{\nabla}^2 \right ) \right ) s - u \ , { \widetilde s}^2 \ , s + d \ , u \ , s \int^t s(t ' ) \biggr ] \ . \label{ii6dipf}\ ] ] for the field theory thus defined , one may take the _
quasistatic limit _ by introducing the fields @xmath754 for @xmath336 , the action ( [ ii6dipf ] ) thus becomes @xmath755 = \int \ !
\d^dx \ , { \tilde \varphi } \bigl [ r - \vec{\nabla}^2 - u \left ( { \tilde \varphi } - \varphi \right ) \bigr ] \varphi \ , \label{ii6qsta}\ ] ] which is known to describe the critical exponents of _ isotropic percolation _ @xcite . an isotropic percolation cluster is shown in fig .
[ isopcl ] , to be contrasted with the anisotropic scaling evident in fig .
[ dirpcl](b ) .
the upper critical dimension of isotropic percolation is @xmath756 , and an explicit calculation , with the diagrams of fig .
[ dploop ] , but involving the static propagators @xmath757 , yields the following critical exponents for isotropic percolation , to first order in @xmath758 , @xmath759 in order to calculate the dynamic critical exponent for this _ dynamic isotropic percolation ( dip ) _ universality class , we must return to the full action ( [ ii6dipf ] ) . once again , with the diagrams of fig .
[ dploop ] , but now involving a temporally nonlocal three - point vertex , one then arrives at @xmath760 for a variety of two - loop results , the reader is referred to ref .
it is also possible to describe the _ crossover _ from isotropic to directed percolation within this field - theoretic framework @xcite .
let us next consider _ multi - species _ variants of directed percolation processes , which can be obtained in the particle language by coupling the dp reactions @xmath761 , @xmath762 via processes of the form @xmath763 ( with @xmath764 ) ; or directly by the corresponding generalisation within the langevin representation with @xmath765 , @xmath766 \right ) s_i + \zeta_i \ , \quad r_i[s_i ] = r_i + \sum_j g_{ij } \
, s_j + \ldots \ , \label{ii6msdp } \\ & & \left\langle \zeta_i(\vec{x},t ) \zeta_j(\vec{x}',t ' ) \right\rangle = 2 s_i n_i[s_i ] \ , \delta(\vec{x } - \vec{x } ' ) \ , \delta(t - t ' ) \ , \delta_{ij } \ , , \
n_i[s_i ] = u_i + \ldots \qquad\ \label{ii6msdn}\end{aligned}\ ] ] the ensuing renormalisation factors turn out to be precisely as for single - species dp , and consequently the _ generical _ critical behaviour even in such multi - species systems is governed by the dp universality class @xcite .
for example , the predator extinction threshold for the stochastic lotka volterra system mentioned in sec .
[ ssec : ii1 ] is characterised by the dp exponents as well @xcite . but these reactions also _
@xmath767 , causing additional terms @xmath768 in eqs .
( [ ii6msdp ] ) .
asymptotically , the inter - species couplings become _ unidirectional _ , which allows for the appearance of special _
multicritical points _ when several @xmath769 simultaneously @xcite .
this leads to a _ hierarchy _ of order parameter exponents @xmath770 on the @xmath186th level of a unidirectional cascade , with @xmath771 for the associated _ crossover exponent _ , one can show @xmath772 to _ all _ orders @xcite .
quite analogous features emerge for _ multi - species dip processes _ @xcite . in these lecture notes ,
i have described how stochastic processes can be mapped onto field theory representations , starting either from a mesoscopic langevin equation for the coarse - grained densities of the relevant order parameter fields and conserved quantities , or from a more microsopic master equation for interacting particle systems .
the dynamic renormalisation group method can then be employed to study and characterise the universal scaling behaviour near continuous phase transitions both in and far from thermal equilibrium , and for systems that generically display scale - invariant features . while the critical dynamics near equilibrium phase transitions has been thoroughly investigated experimentally in the past three decades , regrettably such direct experimental verification of the by now considerable amount of theoretical work on _ nonequilibrium _ systems is largely amiss . in this respect
, applications of the expertise gained in the nonequilibrium statistical mechanics of complex cooperative behaviour to biological systems might prove fruitful and constitutes a promising venture .
one must bear in mind , however , that nonuniversal features are often crucial for the relevant questions in biology . in part , the lack of clearcut experimental evidence may be due to the fact that asymptotic universal properties are perhaps less prominent in accessible nonequilibrium systems , owing to long crossover times .
yet fluctuations do tend to play a more important role in systems that are driven away from thermal equilibrium , and the concept of universality classes , despite the undoubtedly much increased richness in dynamical systems , should still be useful .
for example , we have seen that the directed percolation universality class quite generically describes the critical properties of phase transitions from active to inactive , absorbing states , which abound in nature .
the few exceptions to this rule either require the coupling to another conserved mode @xcite ; the presence , on a mesoscopic level , of additional symmetries that preclude the spontaneous decay @xmath501 as in the so - called _ parity - conserving ( pc ) _ universality class , represented by branching and annihilating random walks @xmath773 with @xmath312 _ even _ , and @xmath522 @xcite ( for recent developments based on nonperturbative rg approaches , see ref .
@xcite ) ; or the absence of any first - order reactions , as in the ( by now rather notorious ) _ pair contact process with diffusion ( pcpd ) _
@xcite , which has so far eluded a successful field - theoretic treatment @xcite . a possible explanation for the fact that dp exponents have not been measured ubiquitously ( yet ) could be the instability towards _ quenched disorder _ in the reaction rates @xcite . in reaction
diffusion systems , a complete classification of the scaling properties in _ multi - species _
systems remains incomplete , aside from pair annihilation and dp - like processes , and still constitutes a quite formidable program ( for a recent overview over the present situation from a field - theoretic viewpoint , see ref .
this is even more evident for nonequilibrium systems in general , even when maintained in driven _
steady states_. field - theoretic methods and the dynamic renormalisation group represent powerful tools that i believe will continue to crucially complement exact solutions ( usually of one - dimensional models ) , other approximative approaches , and computer simulations , in our quest to further elucidate the intriguing cooperative behaviour of strongly interacting and fluctuating many - particle systems . this work has been supported in part by the u.s . national science foundation through grant nsf dmr-0308548 .
i am indebted to many colleagues , students , and friends , from and with whom i had the pleasure to learn and research the material presented here ; specifically i would like to mention vamsi akkineni , michael bulenda , john cardy , olivier deloubrire , daniel fisher , reinhard folk , erwin frey , ivan georgiev , yadin goldschmidt , peter grassberger , henk hilhorst , haye hinrichsen , martin howard , terry hwa , hannes janssen , bernhard kaufmann , mauro mobilia , david nelson , beth reid , zoltn rcz , jaime santos , beate schmittmann , franz schwabl , steffen trimper , ben vollmayr - lee , mark washenberger , fred van wijland , and royce zia .
lastly , i would like to thank the organisers of the very enjoyable luxembourg summer school on _ ageing and the glass transition _ for their kind invitation , and my colleagues at the laboratoire de physique thorique , universit de paris - sud orsay , france and at the rudolf peierls centre for theoretical physics , university of oxford , u.k . ,
where these lecture notes were conceived and written , for their warm hospitality . h.k .
janssen : field - theoretic methods applied to critical dynamics . in : _ dynamical critical phenomena and related topics , lecture notes in physics _ , vol .
104 , ed by c.p .
enz ( springer , heidelberg 1979 ) , pp .
26 - 47 u.c .
tuber : _ critical dynamics : a field theory approach to equilibrium and non - equilibrium scaling behavior _ , in preparation
( to be published at cambridge university press , cambridge ) ; for completed chapters , see : + ` http://www.phys.vt.edu/\,\widetilde{}\,tauber/utaeuber.html ` h.k .
janssen : on the renormalized field theory of nonlinear critical relaxation . in : _ from phase transitions to chaos _
, ed by g. gyrgyi , i. kondor , l. sasvri , and t. tl ( world scientific , singapore 1992 ) , pp . | it is explained how field - theoretic methods and the dynamic renormalisation group ( rg ) can be applied to study the universal scaling properties of systems that either undergo a continuous phase transition or display generic scale invariance , both near and far from thermal equilibrium .
part 1 introduces the response functional field theory representation of ( nonlinear ) langevin equations .
the rg is employed to compute the scaling exponents for several universality classes governing the critical dynamics near second - order phase transitions in equilibrium .
the effects of reversible mode - coupling terms , quenching from random initial conditions to the critical point , and violating the detailed balance constraints are briefly discussed .
it is shown how the same formalism can be applied to nonequilibrium systems such as driven diffusive lattice gases .
part 2 describes how the master equation for stochastic particle reaction processes can be mapped onto a field theory action .
the rg is then used to analyse simple diffusion - limited annihilation reactions as well as generic continuous transitions from active to inactive , absorbing states , which are characterised by the power laws of ( critical ) directed percolation .
certain other important universality classes are mentioned , and some open issues are listed . |
in a simple quark spectator model , the lifetimes of all @xmath1 hadrons are equal , independent of the flavor of the light quark(s ) bound to the @xmath7 quark .
however , significant non - spectator effects enter in the third order , @xmath8 , and higher order terms in the heavy quark expansion ( hqe ) @xcite of the decay width , leading to the predicted lifetime hierarchy : @xmath9 .
especially , pauli interference terms increase @xmath10 by 5% and @xmath11 by 3% and weak annihilation and exchange reduce @xmath11 by 7% .
recent theoretical calculations predict @xmath12 , @xmath13 and @xmath14 @xcite .
the experimental world averages for these ratios are @xmath15 , @xmath16 and @xmath17 , respectively @xcite .
while @xmath18 and @xmath19 lifetimes are measured precisely at the @xmath1-factories , the experimental uncertainty on @xmath2 lifetime is far higher than the theory uncertainty .
prior to 2006 world average of @xmath20 was small compared to the 3rd order hqe prediction , while the value measured by cdf in 2006 @xcite is precise but larger than previous results .
the doubly heavy @xmath3 meson decay is an interesting laboratory to probe heavy quark dynamics , where both @xmath7 and @xmath21 quarks can decay weakly as well as annihilate . due to significant contributions from @xmath21 quark decay and annihilation process
, the @xmath3 lifetime is expected to be shorter than the light @xmath1 mesons .
current theory predictions for @xmath22 range from 0.47 to 0.59 ps @xcite . in this
proceeding we present world best measurements of @xmath2 , @xmath5 and @xmath3 lifetimes by cdf which will help constrain the theories .
in addition , we present a measurement of @xmath18 lifetime in a decay time biased data sample using a monte carlo independent method .
it significantly reduces systematic uncertainty on lifetime compared to the traditional monte carlo based method and can provide improvements on future measurements of other @xmath1 hadrons .
the @xmath3 lifetime is measured in @xmath23 , @xmath24 semileptonic decay channels @xcite collected by di - muon trigger in 1.0 fb@xmath0 data . due to the missing neutrino momentum , the observed decay time ( @xmath25 )
is corrected by a @xmath26-factor as follows : @xmath27 where , @xmath28 and @xmath29 , extracted from monte carlo .
about 5.6 million signal @xmath30 events were reconstructed .
the main challenge in the analysis comes from various backgrounds which enter due to using a broad partially reconstructed mass shape .
these include real @xmath30 plus misidentified lepton , random track pair consistent to be @xmath30 plus real lepton , real @xmath30 plus real lepton coming from different @xmath7 quarks , prompt @xmath30 plus lepton and residual conversion in @xmath31 decays .
the decay time distribution shape of all backgrounds are modeled and calibrated carefully both from data and monte carlo .
the lifetime is extracted from an unbinned maximum likelihood fit where the signal and background normalizations are fixed and @xmath22 is the only free parameter .
the measured lifetimes from the electron and muon channels are @xmath32 and @xmath33 and the combined result is @xmath34 which is the world best measurement to date .
figure [ fig : bcctau ] ( left ) shows the combined likelihood as a function of the @xmath35 . combining with the d measurements @xcite
the tevatron weighted average is @xmath36 a comparison of our @xmath3 lifetime result with run i cdf and dresults is shown in figure [ fig : bcctau ] ( right ) .
a good agreement is seen between all the measurements .
the @xmath2 lifetime is measured in the flavor - specific @xmath37 decays @xcite , with @xmath38 collected using cdf displaced vertex trigger in a 1.3 fb@xmath0 data sample .
in addition to about 1100 fully reconstructed @xmath39 events this analysis uses about 2000 partially reconstructed @xmath40 and @xmath41 events to increase the statistics significantly .
the latter type events involve missing particles ( e.g. a @xmath42 in the @xmath43 mode ) which result in mis - reconstructed mass and transverse momentum .
a @xmath26-factor is introduced to correct the lifetime : @xmath44 the @xmath2 lifetime is determined from two sequential fits .
the first is a binned maximum likelihood fit of the invariant mass distribution of the @xmath39 candidates to determine the signal and background composition .
the second is a unbinned maximum likelihood fit of @xmath45 to extract the @xmath2 lifetime where all the other parameters are fixed .
decay time bias due to trigger acceptance is modeled in the final fit by an efficiency curve , extracted from monte carlo .
extensive tests of the fit procedure have been performed on @xmath18 and @xmath19 control samples which yield lifetime results in good agreement with the corresponding world average values .
figure [ fig : bsctau ] ( left ) shows the lifetime fit projection .
we obtain the flavor - specific @xmath2 lifetime @xmath46 this is the current world best measurement in flavor - specific mode .
figure [ fig : bsctau ] ( right ) compares this result with all published results from flavor - specific channels .
a good overall agreement with the pdg 2007 @xcite result is seen .
the @xmath5 lifetime is measured in @xmath47 fully reconstructed decays @xcite obtained from the cdf displaced vertex trigger where @xmath48 is reconstructed from its decay to a proton , kaon and a pion .
about 2900 signal events are reconstructed from a data sample of 1.1 fb@xmath0 .
a blind analysis is performed where all analysis procedures are finalized with the signal region ( @xmath49 } \rm{gev}/c^2 $ ] in data blinded . as in section [ bslife ]
a two - step fit of invariant mass and lifetime is performed to extract the @xmath5 lifetime .
as before , the trigger efficiency curve , obtained from monte carlo , is introduced in the lifetime fit to model decay time bias .
the fit procedure is validated using input lifetime ( @xmath50 ) values in monte carlo between 325 and 500 @xmath51 m .
figure [ fig : lbctau ] ( left ) shows the lifetime fit projection .
we measure the @xmath5 lifetime @xmath52 this is the current world best measurement . using world average @xmath19 lifetime @xcite
we obtain @xmath53 the leading sources of systematic error to this measurement come from modeling of the displaced vertex trigger , @xmath54 dalitz structure and decay time distribution of combinatorial background .
figure [ fig : lbctau ] ( right ) compares this result with the current world average and all measurements contributing to it , where a good agreement is seen .
this result is also compatible with the hqe prediction of the lifetime ratio , @xmath20 .
the cdf displaced vertex trigger provides heavy - flavor enriched data sample crucial for many @xmath1 physics analyses , two of which are presented in sections [ bslife ] and [ lblife ] .
the decay time distributions from this sample are biased which @xmath2 and @xmath5 lifetime analyses correct for , employing monte carlo based efficiency functions .
this method , however , serves as the dominant source of systematic error in both these analyses .
a monte carlo independent method has been developed which exploits the decay kinematics in data to correct for the trigger bias on a per event basis .
it completely avoids monte carlo modeling of the trigger effect and thus promises to improve uncertainty involved in future lifetime measurements in trigger biased samples . as a proof of principle
, we present a measurement of @xmath18 lifetime measured in the fully reconstructed @xmath55 decays @xcite .
about 24200 signal events are reconstructed in 1 fb@xmath0 data .
the cdf displaced vertex trigger requires a pair of tracks with large impact parameters .
the monte carlo independent method `` slides '' the events along the @xmath1 meson flight direction and evaluates trigger acceptance as a function of proper decay length .
decays ( right ) .
[ fig : buctau],title="fig:",width=321 ] decays ( right ) .
[ fig : buctau],title="fig:",width=264 ] the resulting averaged trigger acceptance functions for signal and sideband regions are shown in figure [ fig : buctau ] ( left ) . since the per event acceptance is a function and not a scalar variable it greatly complicates the fit . to circumvent this ,
a fisher discriminant analysis is performed to transform the acceptance into a variable which is easier to handle .
the following three step method is adopted to extract the @xmath18 lifetime .
firstly the mass distribution is fitted which defines the signal and sideband regions and provides the sample composition .
secondly , using the signal and sideband regions the fisher variable is computed from the trigger acceptance .
finally a maximum likelihood fit is performed on the lifetime and the signal dependent fisher variable to extract the lifetime .
figure [ fig : buctau ] ( right ) shows the lifetime fit projection .
we measure the @xmath18 lifetime @xmath56 this is in good agreement with the current world average of @xmath57 @xcite .
the small systematic error of this measurement makes the monte carlo independent method promising for future lifetime measurements of , e.g. @xmath2 and @xmath5 hadrons .
the dominant sources of systematic uncertainties are impact parameter and transverse momentum dependence of track finding efficiency and correlation betwen mass and decay time in the background .
using data samples in excess of 1 fb@xmath0 collected by the cdf detector we present world best measurements of @xmath3 , @xmath2 and @xmath5 hadron lifetimes .
they are in good agreement with the corresponding experimental world averages and hqe predictions .
a monte carlo independent method is used to measure the @xmath18 lifetime in a trigger biased data sample which agrees with the current world average and is expected to be used in future measurements of other @xmath1 hadron lifetimes .
it is to be noted that all the measurements presented in this proceeding are from about 1/4@xmath58 of the current cdf dataset .
we expect significant improvements to these results in the near future as more data are analyzed . | using data samples in excess of 1 fb@xmath0 collected by the cdf ii detector we present several world s best measurements of @xmath1 hadron lifetimes .
they include @xmath2 meson lifetime in a flavor - specific decay mode , combining fully and partially reconstructed hadronic decays , @xmath3 meson lifetimes in semileptonic @xmath4 decays and @xmath5 baryon lifetime in @xmath6 fully - reconstructed decays .
in addition , we introduce a monte carlo independent technique for measuring @xmath1 hadron lifetimes in data samples biased by displaced vertex triggers . |
the quantum hall effect , discovered in 1980 by klaus von klitzing,@xcite allows one to determine the quantum resistance standard in terms of the electron charge and plank s constant with a spectacular accuracy .
however , the level of precision necessary for metrology applications ( a few parts per billion ) requires for conventional quasi - two - dimensional semiconductor systems the use of ultra - low temperatures and high magnetic fields.@xcite the discovery of graphene@xcite in 2004 led to a revival of interest in quantum hall effect physics.@xcite the energy of landau levels ( lls ) in graphene is given by @xmath0 where @xmath1 is the fermi velocity , @xmath2 is the magnetic length , @xmath3 is the magnitude of the applied magnetic field , @xmath4 is the ll number , and the `` @xmath5 '' signs correspond to the electrons and holes respectively .
@xcite for @xmath6 the separation between the zero and the first ll in graphene is @xmath7 ( in comparison , @xmath8 in conventional quasi - two - dimensional semiconductor systems ) . therefore , the quantum hall effect in graphene can be observed even at room temperatures in experimentally attainable magnetic fields .
indeed , the room - temperature quantum hall effect in graphene was first observed in 2007;@xcite however , extremely high magnetic fields ( up to @xmath9 ) were needed .
it has been understood since the 1930 s that the longitudinal conductivity of metals in a quantizing magnetic field increases with increasing electron scattering.@xcite to develop a graphene - based quantum hall standard of resistance,@xcite which would work at elevated temperatures and in moderate magnetic fields , it is essential to examine the contributions from different scattering processes to the longitudinal conductivity @xmath10 in the quantum hall regime ( qhr ) , @xmath11 , as @xmath10 provides the major correction to the quantized value of the hall resistance . in conventional semiconductor systems
the quantum hall effect is observed at liquid helium temperatures only , and the value of @xmath10 is governed by scattering on disorder . in the qhr , @xmath10 depends exponentially on the energy separation between the fermi level and the nearest ll .
the longitudinal conductivity prefactor has a theoretically - predicted value of @xmath12 in the presence of short - range disorder and @xmath13 in the presence of long - range disorder.@xcite it was shown that the contribution from phonon scattering is negligibly small .
the exceptions are exotic cases such as magnetoroton dissociation , in which the whole effect arises from scattering on phonons,@xcite or an enhancement of phonon - induced scattering near the intersection of two lls corresponding to different size - quantization subbands.@xcite however , as we show in this paper , the phonon - scattering mechanism dominates in the high - temperature qhr in graphene , since at @xmath14 ( where @xmath15 is the sound velocity in graphene @xcite ) the energy of an acoustic phonon with a wavevector comparable to the inverse magnetic length is much smaller than the temperature ; therefore , the number of such phonons increases drastically . in this paper
we restrict our consideration to electron scattering induced by the interaction with intrinsic in - plane phonons only , neglecting the effects of disorder as well as electron - electron interactions .
our calculations result in the lower estimate of the longitudinal conductivity @xmath10 in graphene in the qhr , as we do not study electron interactions with various other types of phonons such as out - of - plane flexural phonons , which are present in suspended samples,@xcite or bulk acoustic phonons in a substrate , which interact with electrons in graphene on polar substrates such as boron nitride @xcite or silicon carbide via piezoelectric coupling.@xcite
the wave function of a charged carrier ( electron or hole ) in graphene subjected to a magnetic field @xmath16 normal to the graphene plane is given in the landau gauge , @xmath17 , by the following expression:@xcite @xmath18 \left[{h_{n}\left [ \left(x - x_{0}\right ) / l_{b } \right ] \atop \pm \sqrt{2 n } h_{n-1 } \left [ \left(x - x_{0}\right ) / l_{b } \right]}\right ] \frac { \exp \left(ik_{y } y\right ) } { \sqrt{l_{y } } } \mbox{,}\ ] ] where @xmath19 and @xmath20 , @xmath21 is the guiding centre coordinate , @xmath22 is the electron wavevector @xmath23-component , @xmath24 are the hermite polynomials , @xmath25 and @xmath26 are the graphene sample dimensions , the `` @xmath5 '' sign corresponds to electrons and holes , and the spin and valley indeces are omitted . here
the @xmath27-axis has been chosen in the direction of the applied magnetic field and the @xmath28-axis and @xmath23-axis are in the graphene plane . the wave function given by eq.([2componentwf ] )
is defined for one dirac point only and its two components correspond to graphene s two sublattices . in what follows all results will be obtained for electrons only .
calculations for the non - equivalent dirac point and holes can be easily repeated in a similar fashion . from here on we will omit the `` @xmath5 '' sign in order to simplify notation .
we account for the valley and spin degeneracy by multiplying the final result for @xmath10 by the factor of four at the final stage of calculations .
electron scattering on phonons leads to a change in the @xmath22 component of the electron wavevector , which results in a change to the electron guiding centre coordinate @xmath29 .
electron transitions between lls due to one phonon scattering are suppressed due to the large energy gaps between lls in graphene in the qhr .
indeed , the number of phonons with the energy required for such transitions is @xmath30 ( @xmath31 is the energy gap between two different lls in graphene ) , which is very small in the qhr .
inter - ll scattering on acoustic phonons has an additional exponentially strong suppression in the matrix element of transition which is markedly different from the case of conventional semiconductor systems as discussed in the appendix .
it is also evident that in the qhr only scattering on acoustic phonons can provide a noticeable contribution to the longitudinal conductivity @xmath10 .
the optical phonon energy in graphene corresponds to the temperature range @xmath32 @xcite which leads to very small optical phonon occupation numbers at room temperatures and below . in this work ,
we are interested in phonon - induced equilibrium longitudinal conductivity , whereas , magneto - phonon resonance associated with optical phonons has been studied in ref .
one - phonon scattering within the same ll is ineffective because of the small width of the lls , which can be arguably achieved in pristine graphene samples .
interestingly , even in heavily disordered samples the broadening of lls was found to shrink in several particular cases .
@xcite for our calculations we assume an infinitely narrow band of delocalized states in the middle of each ll with a vanishing density of states in between lls . in this limit one - phonon scattering within one ll is forbidden by energy and momentum conservation . the next order process to consider is two - phonon scattering through virtual states with no change in the electron initial and final ll numbers but with the change in its in - plane momentum ( or the guiding centre coordinate ) .
we calculate the longitudinal conductivity @xmath10 at the @xmath4th ll due to two - phonon scattering using the generalized form of the einstein relation:@xcite @xmath33 @xmath34 where @xmath35 + 1 \right \ } ^{-1}$ ] is the ll filling factor , @xmath36 is the ll energy defined by eq.([landaulevels ] ) , @xmath37 is the fermi energy , @xmath22 and @xmath38 are the @xmath23-components of the electron wavevector before and after scattering , @xmath39 is the diffusion coefficient , and @xmath40 is the probability of scattering .
two - phonon scattering in graphene in the qhr is possible through two different virtual intermediate states : a phonon with wavevector @xmath41 is first emitted or a phonon with wavevector @xmath42 is first absorbed .
transitions changing the electron ll number in the intermediate states are strongly suppressed due to small values of corresponding matrix elements and the presence of large denominators in the expression for @xmath40 ( see appendix ) .
therefore , in what follows we consider only transitions with no change in the electron ll number .
then , the probability of two - phonon scattering is given by fermi s golden rule with the matrix element @xmath43 where @xmath44 ^{-1}$ ] is the phonon occupation number , @xmath45 is the acoustic phonon frequency , and @xmath46 is the electron - phonon coupling operator .
operators describing electron scattering on intrinsic longitudinal ( la ) and transverse ( ta ) acoustic phonons in graphene are @xcite @xmath47 @xmath48 where @xmath49 , @xmath50 is the graphene 2d mass density , @xmath51 is the position vector in the graphene plane , and @xmath52 is the angle between the phonon wavevector @xmath53 and the @xmath28-axis . the diagonal matrix elements in eqs.([efinteractionla])-([efinteractionta ] ) describe electron coupling to the phonon - created deformation potential , and the off - diagonal matrix elements originate from the phonon - induced bond - length modulations , which effect hopping amplitudes between two neighbouring sites .
the corresponding coupling constants were estimated as @xmath54 and @xmath55.@xcite note that the deformation potential couples electrons with la phonons only .
furthermore , since the second component of the electron wave function defined by eq.([2componentwf ] ) vanishes for the zero ll , electrons in this ll do not interact with ta phonons . substituting the electron - phonon scattering operators given by eqs.([efinteractionla])-([efinteractionta ] ) into eq.([m_element_1 ] ) yields @xmath56 where @xmath57 , @xmath58 are the indices introduced to separate the contributions to the longitudinal conductivity @xmath10 from la and ta phonons and from the two different scattering mechanisms discussed above , @xmath59 is the generalized electron - phonon coupling constant with the values @xmath60 , @xmath61 , @xmath62 , @xmath63 , and the matrix elements @xmath64 are given by @xmath65 \delta_{k_{y}',k_{y } \pm q_{y } } \mbox{,}\ ] ] @xmath66 \exp \left(\pm \beta - \alpha\right ) l_{n-1}^{1 } \left ( \alpha \right ) \delta_{k_{y } ' , k_{y } \pm q_{y } } \mbox{,}\end{gathered}\ ] ] @xmath67 \exp \left(\pm \beta - \alpha\right ) l_{n-1}^{1 } \left ( \alpha \right ) \delta_{k_{y } ' , k_{y } \pm q_{y } } \mbox{.}\end{gathered}\ ] ] in eqs.([m_la_d])-([m_ta_h ] ) , @xmath68 $ ] , @xmath69 , @xmath70 , the `` @xmath5 '' sign refers to emitted and absorbed phonons , @xmath71 and @xmath72 are the laguerre polynomials , @xmath73 is the @xmath74 identity matrix , and @xmath75 , @xmath76 are given by @xmath77 eq.([m_la_d ] ) is also valid for the zero ll when substituting @xmath78 and @xmath79 .
note the @xmath80 factor in eqs.([m_la_h])-([m_ta_h ] ) which suppresses the off - diagonal contribution to @xmath10 in higher lls .
summation over all possible values of @xmath81 in eq.([m_element_2 ] ) results in the following expressions for two - phonon scattering matrix elements @xmath82^{2 } \left [ l_{n}^{0 } \left ( l_{b}^{2 } q^{2 } / 2 \right ) + l_{n-1}^{0 } \left ( l_{b}^{2 } q^{2 } / 2 \right ) \right]^{4 } \\ \times \exp \left ( - l_{b}^{2 } q^{2 } \right ) \sin ^{2 } \left [ l_{b}^{2 } \left ( q_{y}^{+ } q_{x}^{- } - q_{y}^{- } q_{x}^{+ } \right ) / 2 \right ] \delta_{k_{y } ' , k_{y } + q_{y}^{- } - q_{y}^{+ } } \mbox{,}\end{gathered}\ ] ] @xmath83^{2 } n^{-2 } \left [ l_{n-1}^{1 } \left ( l_{b}^{2 } q^{2 } / 2 \right ) \right]^{4 } \exp \left ( - l_{b}^{2 } q^{2 } \right ) \\ \times l_{b}^{4 } \left ( q_{x}^{+ } \sin{2 \varphi^{+ } } + q_{y}^{+ } \cos{2 \varphi^{+ } } \right)^{2 } \left ( q_{x}^{- } \sin{2 \varphi^{- } } + q_{y}^{- } \cos{2 \varphi^{- } } \right)^{2 } \\ \times \sin ^{2 } \left [ l_{b}^{2 } \left ( q_{y}^{+ } q_{x}^{- } - q_{y}^{- } q_{x}^{+ } \right ) / 2 \right ] \delta_{k_{y } ' , k_{y } + q_{y}^{- } - q_{y}^{+ } } \mbox{,}\end{gathered}\ ] ] @xmath84^{2 } n^{-2 } \left [ l_{n-1}^{1 } \left ( l_{b}^{2 } q^{2 } / 2 \right ) \right]^{4 } \exp \left ( - l_{b}^{2 } q^{2 } \right ) \\
\times l_{b}^{4 } \left ( q_{x}^{+ } \cos{2 \varphi^{+ } } - q_{y}^{+ } \sin{2 \varphi^{+ } } \right)^{2 } \left ( q_{x}^{- } \cos{2 \varphi^{- } } - q_{y}^{- } \sin{2 \varphi^{- } } \right)^{2 }
\\ \times \sin ^{2 } \left [ l_{b}^{2 } \left ( q_{y}^{+ } q_{x}^{- } - q_{y}^{- } q_{x}^{+ } \right ) / 2 \right ] \delta_{k_{y } ' , k_{y } + q_{y}^{- } - q_{y}^{+ } } \mbox{.}\end{gathered}\ ] ] substituting the calculated probability of the two - phonon scattering @xmath85 into eq.([diffcoefficient ] ) and performing summation over @xmath38 as well as integration over all possible values of @xmath41 and @xmath42 yield the following result for the longitudinal conductivity @xmath86 where @xmath87 \left [ l_{n}^{0 } \left ( l_{b}^{2 } q^{2}/2 \right ) + l_{n-1}^{0 } \left ( l_{b}^{2 } q^{2 } / 2 \right ) \right ] ^{4 } dq \mbox{,}\end{gathered}\ ] ] @xmath88 \left [ l^{1}_{n-1 } \left ( l_{b}^{2 } q^{2 } / 2 \right ) \right ] ^{4 } dq \mbox{.}\end{gathered}\ ] ] here @xmath89 is the bessel function of the first kind .
the expressions for @xmath90 were multiplied by a factor of four to account for the valley and spin degeneracy . from eqs.([sigma_d])-([sigma_h ] ) it is evident that @xmath91 due to the relatively small value of @xmath92 comparing to @xmath93 @xcite and the small @xmath94 factor contained in eq.([sigma_h ] ) .
there are two distinctive temperature limits . in the low - temperature limit , @xmath95 , the main value of the integrals in eqs.([sigma_d])-([sigma_h
] ) is formed by @xmath96 and @xmath97 can be considered as a small parameter .
expanding the integrand into power series in @xmath97 and taking into account only the lowest power term yields the following expression for @xmath90 in the low - temperature limit @xmath98 \left ( t/ t_{l_{b } } \right ) ^{8 } \mbox{,}\ ] ] @xmath99 \left ( 5528 \pi ^{11}/ 1365 \right ) \left [ { g}_{h}^{4 } / \left ( l_{b}^{2 } \rho^{2 } \hbar^{2 } s^{6 } \right ) \right ] \left ( t/ t_{l_{b } } \right ) ^{12 } \mbox{.}\ ] ] in eq.([sigma_h_lt ] ) the coefficients @xmath100 are defined by @xmath101^{4}= \sum \limits_{j=0}^{4n-4 } a_{j}^{n-1 } x^{j}$ ] , where @xmath102 . for the zeroth ll , @xmath103 .
note that @xmath104 in eq.([sigma_d_lt ] ) does not depend on the ll number .
the temperature dependencies given by eqs.([sigma_d_lt],[sigma_h_lt ] ) are different from the case of conventional semiconductor heterostructures for both two - phonon scattering @xcite and phonon - assisted hopping conductivity .
@xcite in eq.([sigma_d_lt ] ) , which corresponds to the deformation potential scattering mechanism , the lower power in the temperature dependence of mobility compared to that obtained in refs .
[ ] is due to phonons in graphene being two - dimensional .
as was mentioned above , the contribution to longitudinal conductivity at higher lls given by eq.([sigma_h_lt ] ) is a distinctive feature of graphene without analogy in semiconductor systems . in the more interesting high - temperature limit , @xmath105 , the main value of the integrals in eqs.([sigma_d])-([sigma_h ] )
is formed by @xmath106 , and @xmath107 can be considered as a small parameter . by expanding the integrand into power series in @xmath107 and taking into account only the lowest power term we obtain the following result for @xmath90 in the high - temperature limit @xmath108 where @xmath109\mbox{,}\ ] ] and @xmath110 \left [ { g}_{h}^{4 } / \left ( l_{b}^{2 } \rho^{2 } \hbar^{2 } s^{6 } \right ) \right ] \left ( t / t_{l_{b } } \right ) \mbox{,}\ ] ] where @xmath111 \mbox{.}\ ] ] in eqs.([phi_d])-([phi_h ] )
, @xmath112 is the gamma function and @xmath113 is the hypergeometric function . in eq.([sigma_h_ht ] ) , @xmath70 , since in the zeroth ll @xmath103 .
the coefficients @xmath114 are the same as in eq.([sigma_h_lt ] ) and the coefficients @xmath115 are defined by @xmath116^{4}= \sum \limits_{j=0}^{4n } b_{j}^{n } x^{j}$ ] .
in fig.[fig_sigma ] we plot @xmath117 obtained by numerical integration of eqs.([sigma_d])-([sigma_h ] ) from @xmath118 to @xmath119 for @xmath120 ( @xmath121 ) . as a function of temperature for @xmath122 .
( b ) expanded view in the @xmath123 temperature range.,title="fig : " ] as a function of temperature for @xmath122 .
( b ) expanded view in the @xmath123 temperature range.,title="fig : " ] it can be seen from fig.[fig_sigma](a ) that for @xmath124 , @xmath125 linearly increases with temperature in line with eqs.([sigma_d_ht])-([sigma_h_ht ] ) . from fig.[fig_sigma](b ) one can see the change of the @xmath126 temperature dependence from the high - power to the linear law occurring around @xmath127
. substituting the numerical values of all the constants into eqs.([sigma_d_ht ] ) and ( [ sigma_h_ht ] ) results in the following simplified expression for @xmath126 at elevated temperatures @xmath128 where @xmath129 has the following values for the six lowest lls : @xmath130 , @xmath131 , @xmath132 , @xmath133 , @xmath134 , @xmath135 . here
we used the unscreened value of the deformation potential , which is arguably appropriate in the strong magnetic field as the electron motion is quantized and the screening is suppressed .
there is a natural question of the applicability of the lowest allowed ( in our case second ) order of perturbation theory in electron - phonon interaction at elevated temperatures .
this problem was studied in detail in ref .
[ ] for conventional semiconductor quantum wells and bulk acoustic phonon scattering . to compare our results with the analysis provided in ref .
[ ] it is necessary to express them in terms of the diffusion coefficient @xmath136 in ref .
[ ] the diffusion coefficient for @xmath105 in the lowest ( second ) order of perturbation theory is written as @xmath137 .
thus , there is a simple connection between the dimensionless electron - phonon interaction constant @xmath138 and our dimensionless constants @xmath139 .
namely , @xmath140 according to ref .
[ ] higher orders of perturbation theory can be neglected for @xmath141 .
unlike the case of conventional semiconductors , for which @xmath142 , for graphene @xmath143 ; therefore , the temperature @xmath144 defining the validity of perturbation theory becomes independent of magnetic field .
for the zero landau level and for the parameters used in our calculations , @xmath145 . above this temperature
the phonon - induced mobility is expected to change from linear to sub - linear temperature dependence and eventually to saturate . for higher landau levels
the perturbation theory cut - off temperature @xmath144 well exceeds @xmath146 and the lowest order perturbation analysis is fully valid for ambient conditions . in conclusion
, we obtained the value of the longitudinal conductivity in graphene in the quantum hall regime due to two - phonon scattering at elevated temperatures which is comparable to the disorder - induced longitudinal magneto - conductivity in conventional semiconductor heterostructures.@xcite the predicted distinctive temperature and magnetic field dependence of the phonon scattering contribution to the pre - exponential factor in @xmath10 given by eq .
( [ sigma_lineart ] ) can be easily separated from the temperature- and field - independent contribution caused by disorder when analysing experimental data .
this should allow the parameters of electron - phonon interaction in graphene to be extracted with enhanced accuracy .
we thank charles downing for a critical reading of the manuscript .
a.m.a . is grateful to daniil alexeev for fruitful discussions .
this work was supported by the eu h2020 rise project coexan ( grant no .
h2020 - 644076 ) , eu fp7 itn notedev ( grant no .
fp7 - 607521 ) , fp7 irses projects cantor ( grant no .
fp7 - 612285 ) , qocan ( grant no .
fp7 - 316432 ) , and internom ( grant no .
fp7 - 612624 ) .
r.r.h . acknowledges financial support from urco ( grant no .
15 f u / s 1tay13 - 1tay14 ) and research links travel grant by the british council newton fund .
the probability of two - phonon scattering through a virtual intermediate state is calculated using fermi s golden rule as given below @xmath147 where @xmath148 and @xmath149 here @xmath150 , @xmath151 , and @xmath152 are the energies of the electron lls in the initial , intermediate and final states . clearly , the transitions changing the electron ll number in the intermediate states are suppressed compared to the transitions conserving the ll number due to the presence of large denominators in eq .
( [ m_element_suppl ] ) .
notably , single - phonon scattering on acoustic phonons with a change in the ll number is also suppressed because of the very small value of the corresponding matrix element @xmath153 due to energy conservation , @xmath154 , which results in the following estimate @xmath155 \mbox{.}\ ] ] interestingly , unlike the case of a conventional quasi - two - dimensional semiconductor system,@xcite the exponential factor in eq .
( [ exponent ] ) does not depend on the magnetic field .
instead , it has a non - trivial dependence on the difference between the ll numbers of the two involved levels and becomes non - vanishing for very high adjacent lls .
however , these high levels are not relevant for the high - temperature quantum hall effect , which is the subject of our interest ; whereas , for @xmath156 and @xmath157 , @xmath158 leading to a complete suppression of inter - ll transitions . | one of the most distinctive features of graphene is its huge inter - landau - level splitting in experimentally attainable magnetic fields which results in the room - temperature quantum hall effect . in this paper
we calculate the longitudinal conductivity induced by two - phonon scattering in graphene in a quantizing magnetic field at elevated temperatures .
it is concluded that the purely phonon - induced scattering , negligible for conventional semiconductor heterostructures under quantum hall conditions , becomes comparable to the disorder - induced contribution to the dissipative conductivity of graphene in the quantum hall regime . |
the structural changes between @xmath2 and @xmath3 of gallium isotopes recently attracted much experimental attention . at argonne national laboratory , stefanescu _ et .
@xcite have populated odd - a @xmath4ga isotopes in deep - inelastic reactions .
more recently , in the coulomb excitation experiment at rex - isolde , the existence of a @xmath5 , @xmath6 ground - state doublet has been proposed in @xmath7ga @xcite . for odd - even ga isotopes
nuclear spins and moments has been reported in @xcite . recently
ground - state spins and moments of @xmath8ga isotopes using laser spectroscopy has been reported in @xcite .
the evolution of the @xmath5 and @xmath9 levels in odd - a gallium isotopes are shown in fig . [ f_gaintro ] . except for @xmath7ga and @xmath10ga
all have ground state ( g.s . ) @xmath6 . the first @xmath5 reaches minimum in @xmath7ga where it becomes the ground state and the first @xmath9 start decreasing with @xmath2 onwards and it becomes the ground state in @xmath10ga .
this figure also demonstrate abrupt changes of structure from @xmath2 to @xmath11 .
following our recent shell - model ( sm ) studies for neutron - rich even isotopes of fe @xcite , odd - odd mn isotopes @xcite , and odd - mass @xmath12co isotopes @xcite , in this paper we report large scale shell model calculations for @xmath13ga isotopes .
earlier shell model calculation using pairing plus quadrupole - quadrupole interaction for @xmath14ga isotopes have been reported by yoshinaga _ et .
al . , _ @xcite .
they consider only low - lying negative - parity states in the analysis .
the aim of present study is to analyze recently accumulated experimental data which includes both positive and negative high - spin states on neutron - rich ga isotopes .
further , this work will add more information in the earlier work for odd - even ga isotopes by cheal _
@xcite , where only few excited negative - parity states ( @xmath15 mev ) are studied in @xmath0 space .
present work also include further theoretical development which is proposed in @xcite , by including @xmath16 orbital in the model space .
the present shell model calculations have been carried out in @xmath0 and @xmath17 spaces . in the first set ,
calculations have been performed with two recently derived effective shell model interactions , jun45 and jj44b , that have been proposed for 1@xmath18 , 0@xmath19 , 1@xmath20 and 0@xmath21 single - particle orbits .
jun45 which is recently develop by honma _
@xcite , this realistic interaction based on the bonn - c potential with fitting 400 experimental binding and excitation energy data with mass numbers a=63@xmath2296 .
since the present model space is not sufficient to describe collectivity in these region thus data have been not used while fitting in the middle of the shell along @xmath23 line . for jun45 , the data mostly fitted to develop this interaction closure to @xmath3 ,
thus this interaction is very suitable for shell - model study of neutron rich ga isotopes as @xmath24 .
although this interaction is not successful to explain data for ni and cu isotopes possibly due to missing 0@xmath16 orbit in the present model space .
the jj44b interaction due to brown _ et al . _
@xcite was developed by fitting 600 binding energies and excitation energies with @xmath25 and @xmath26 . instead of 45 as in jun45 , here
30 linear combinations of good @xmath27 two - body matrix elements varied , with the rms deviation of about 250 kev from experiment . for jun45 interaction the single - particle energies
are taken to be -9.8280 , -8.7087 , -7.8388 , and -6.2617 mev for the @xmath18 , @xmath19 , @xmath20 and @xmath21 orbit , respectively .
similarly for jj44b interaction the single - particle energies are taken to be -9.6566 , -9.2859 , -8.2695 , and -5.8944 mev for the @xmath18 , @xmath19 , @xmath20 and @xmath21 orbit , respectively . in the second set of calculations for the @xmath17 space , with @xmath28ca core ( a @xmath29ca core , with eight neutrons
were frozen in the @xmath30 orbital ) , an interaction reported by sorlin _
et al_. in @xcite has been employed .
this interaction , called @xmath1 interaction , was built using @xmath31 two - body matrix elements ( tbme ) from @xcite and @xmath32 tbme from @xcite .
the remaining @xmath33 tbme are taken from @xcite .
the single - particle energies are taken to be 0.0 , 2.0 , 4.0 , 6.5 and 9.0 mev for the @xmath16 , @xmath18 , @xmath20 , @xmath19 , and @xmath21 orbit , respectively .
as here the dimensions of the matrices become very large , a truncation has been imposed .
we used a truncation by allowing up to a total of three particle excitations from the @xmath16 orbital to the upper @xmath34 orbitals for protons and from the upper @xmath31 orbitals to the @xmath21 orbital for neutrons .
all the sm calculations are performed on the 20-node cluster computer at prl using the code antoine @xcite .
stefanescu _ et al .
_ @xcite assigned positive parity states at 2082 , 2941 , and 4028 kev , for 13/2@xmath36 , 17/2@xmath36 , and 21/2@xmath36 respectively . in the same experiment
the negative states 9/2@xmath37 , 13/2@xmath37 , 17/2@xmath37 , and ( 19/2@xmath37 , 21/2@xmath37 ) are proposed at 1498 , 2684 , 3696 , and 4165 kev , respectively . in this paper
they argue that these sequence of levels arises due to coupling of @xmath38 and @xmath39 . in fig .
[ f_71ga ] , experimental data for @xmath35ga are compared with the calculated energy levels using jun45 , jj44b and @xmath1 interactions . the band built on 3/2@xmath37 , is reproduced correctly by jun45 interaction , while levels are compress with jj44b interaction .
these results show that the yrast sequence of levels @xmath6@xmath9@xmath40@xmath41 is connected with strong @xmath42 transitions .
the @xmath43 values for 5/2@xmath37@xmath443/2@xmath37 , 7/2@xmath37@xmath445/2@xmath37 , 11/2@xmath37@xmath447/2@xmath37 , are 6 , 222 and 223 e@xmath45@xmath46 respectively with jun45 and corresponding value for jj44b are 5 , 209 and 383 e@xmath45@xmath46 .
similarly , a band built on 5/2@xmath37 , with jun45 interaction is slightly higher , while jj44b interaction predicted compress states .
the band built on 9/2@xmath36 , with the jun45 interaction , predicted 9/2@xmath36 , 13/2@xmath36 , 17/2@xmath36 , 21/2@xmath36 are at 2478 , 2893 , 3526 and 4979 kev , while its corresponding experimental values are 1493 , 2082 , 2941 and 4028 kev respectively . the calculated positive parity states around 1 mev higher than experimental data with jun45 .
the predicted results show an inadequacy of this interaction to explain positive parity states .
the jj44b interaction predicts 9/2@xmath36 , 13/2@xmath36 , 17/2@xmath36 , 21/2@xmath36 at 1922 , 2855 , 3689 and 4866 kev respectively , which is better than jun45 interaction .
the calculated @xmath43 values are shown in table [ t_be2 ] .
the jun45 interaction predicts , leading configuration of @xmath47@xmath48@xmath49 for the ground state with probability of 29.2% . the calculated occupancy of g@xmath50 orbital is 0.53 , 0.83 by jun45 and jj44b interaction respectively .
states with @xmath51= 5/2@xmath37 , 7/2@xmath37 , 9/2@xmath37 , 11/2@xmath37 , 13/2@xmath37 are dominated with @xmath52 @xmath48 @xmath53 configurations by @xmath1 interaction . while for this interaction positive parity states with @xmath51= 9/2@xmath36 , 13/2@xmath36 , 17/2@xmath36 , 21/2@xmath36 are dominated with @xmath54 @xmath48 @xmath55 configurations .
the 13/2@xmath36 predicted by @xmath1 interaction is within 50 kev with experimental data .
the @xmath1 interaction predict better result for excitation energies in comparison to jun45 and jj44b interactions .
ga compared with shell - model results .
the levels marked with an asterisk ( * ) have been also experimentally found , but not form a band . ]
ga compared with shell - model results .
the levels marked with an asterisk ( * ) have been also experimentally found , but not form a band . ]
cheal _ et al .
_ @xcite assigned 1/2@xmath37 as a ground state , in another experiment recently diriken _
@xcite shown an evidence for a 1/2@xmath37 , 3/2@xmath37 doublet near the ground state , with an excitation energy of the @xmath22 0.8 kev for 3/2@xmath37 . in fig .
[ f_73ga ] , experimental data for @xmath7ga are compared with the calculated energy levels using jun45 , jj44b and @xmath1 interactions .
in contrast to ground state doublet predicted in recent experiment @xcite , the jun45 , jj44b and @xmath1 interactions predict 1/2@xmath37 state at 219 , 91 and 505 kev respectively .
indeed , jun45 predict too high 1/2@xmath37 for @xmath7ga .
similar trend has been also observed with jun45 interaction for cu isotopes where 5/2@xmath37 is too high below @xmath2 and 1/2@xmath37 is too high above @xmath2 .
the band built on 3/2@xmath37 , reproduced correctly by jun45 and jj44b interaction .
the band built on 5/2@xmath37 , is more correctly reproduced by jj44b interaction .
the @xmath43 values for 9/2@xmath37@xmath445/2@xmath37 , 13/2@xmath37@xmath449/2@xmath37 , 17/2@xmath37@xmath4413/2@xmath37 , are 42 , 194 and 4 e@xmath45@xmath46 respectively with jun45 and corresponding value for jj44b are 81 , 270 and 161 e@xmath45@xmath46 .
the results of band built on 7/2@xmath37 , is better for jj44b in comparison to jun45 .
the band built on 9/2@xmath36 , with the jun45 interaction , predicted 9/2@xmath36 , 13/2@xmath36 , 17/2@xmath36 , 21/2@xmath36 are at 2230 , 2749 , 3267 and 4588 kev , while its corresponding experimental values are 1232 , 1814 , 2718 and 3974 kev respectively .
the occupancy of @xmath21 orbital is predicted by jun45 and jj44b interaction as 0.92 and 0.83 respectively . with @xmath1 interaction 13/2@xmath56
become lower than 9/2@xmath56 .
the calculated @xmath43 values obtained with jj44b give reasonable agreement with experimental data .
the experimental excitation energy of 9/2@xmath56 decreases from @xmath35ga to @xmath7ga and again starts rising from @xmath57ga onwards .
indeed , intruder @xmath58 leading to onset of deformation around neutron mid - shell at @xmath11 .
only jun45 interaction predict this trend from @xmath35ga to @xmath7ga . states with @xmath51= 7/2@xmath37 , 9/2@xmath37 , 11/2@xmath37 , 13/2@xmath37 , 15/2@xmath37 , 17/2@xmath37 , 19/2@xmath37 , 21/2@xmath37 are dominated with @xmath54 @xmath48 @xmath59 configurations by jj44b interaction for @xmath7ga .
the overall results of jj44b interaction are better than jun45 and @xmath1 .
ga compared with shell - model results .
the levels marked with an asterisk ( * ) have been also experimentally found , but not form a band . ] in fig.[f_75ga ] , experimental data for @xmath57ga are compared with the calculated energy levels with jun45 , jj44b and @xmath1 interactions . the band built on 3/2@xmath37 , reproduced correctly by jun45 and jj44b interaction .
the band built on 5/2@xmath37 , is more correctly reproduced by jj44b interaction .
the results of band built on 7/2@xmath37 is reasonable with jj44b .
the @xmath43 values for 11/2@xmath37@xmath447/2@xmath37 , 15/2@xmath37@xmath4411/2@xmath37 , 19/2@xmath37@xmath4415/2@xmath37 , are 278 , 109 and 92 e@xmath45@xmath46 respectively with jun45 and corresponding values for jj44b are 388 , 368 and 281 e@xmath45@xmath46 .
the band built on 9/2@xmath36 , with the jun45 interaction , predicted 9/2@xmath36 , 13/2@xmath36 , 17/2@xmath36 , 21/2@xmath36 are at 2206 , 2726 , 3277 and 4576 kev , while its corresponding experimental values are 1510 , 2088 , 2946 and 4148 kev respectively .
the ground state having configuration @xmath60 with probability 14.9 and 22.4% for jun45 and jj44b interaction .
the results of @xmath1 interaction are more compressed .
states with @xmath51= 7/2@xmath37 , 9/2@xmath37 , 11/2@xmath37 , 13/2@xmath37 , 15/2@xmath37 , 17/2@xmath37 , 19/2@xmath37 , 21/2@xmath37 are dominated with @xmath54 @xmath48 @xmath61 configurations by jj44b interaction for @xmath57ga .
the first @xmath62 has @xmath63 ) @xmath48 @xmath61 configuration with jj44b for @xmath57ga . with jj44b
the occupancy of @xmath21 orbital is 0.49 .
the results of jj44b interaction are better than jun45 and @xmath1 .
ga compared with shell - model results .
the levels marked with an asterisk ( * ) have been also experimentally found , but not form a band . ] in fig.[f_77ga ] , experimental data for @xmath64ga are compared with the calculated energy levels with jun45 , jj44b and @xmath1 interactions . the band built on 3/2@xmath37 ,
is reproduced correctly by jun45 interaction , while levels are slightly higher with jj44b interaction .
the @xmath43 values for 7/2@xmath37@xmath443/2@xmath37 , 11/2@xmath37@xmath447/2@xmath37 , are 45 and 265 e@xmath45@xmath46 with jun45 and corresponding value for jj44b is 38 and 310 e@xmath45@xmath46 . the @xmath43 values for 9/2@xmath37@xmath445/2@xmath37 , 13/2@xmath37@xmath449/2@xmath37 , 17/2@xmath37@xmath4413/2@xmath37 , are 135 , 33 and 105 e@xmath45@xmath46 respectively with jun45 and corresponding value for jj44b are 167 , 78 and 101 e@xmath45@xmath46 .
the band built on 9/2@xmath36 is correctly reproduce by jun45 interaction .
the @xmath65 is predicted at 2724 kev by jun45 , while at 2345 kev by jj44b , which is slightly compressed .
the ground state as @xmath6 having configuration @xmath60 with probability 35.7 and 24.5% for jun45 and jj44b interaction respectively . with jj44b interaction
the 1/2@xmath37 is high for @xmath64ga .
similar trend has been also found for cu isotopes , where 1/2@xmath37 bit too high in @xmath66cu .
states with @xmath51= 7/2@xmath37 , 9/2@xmath37 , 11/2@xmath37 , 13/2@xmath37 , 15/2@xmath37 , 17/2@xmath37 , 19/2@xmath37 , 21/2@xmath37 are dominated with @xmath54 @xmath48 @xmath67 configurations by jun45 interaction for @xmath64ga .
the occupancy of g@xmath50 orbital for first @xmath62 state is 0.96 for jun45 , while it is only 0.053 for jj44b interaction .
the results of jun45 and jj44b are reasonable .
the positive parity states predicted by @xmath1 are more compressed .
in fig . 6 the calculated energy levels for even ga isotopes with jun45 , jj44b and @xmath1 interactions are compared with experimentally known data . recently
ground - state spins and moments for even ga isotopes with a=72,74,76,78 has been reported in @xcite . for @xmath68ga
all the three interactions predict different g.s .
the first 3@xmath37 is predicted at 200 , 131 and 321 kev by jun45 , jj44b and @xmath1 interaction respectively . in @xmath69ga
the jun45 interaction predict 3@xmath37 as a g.s . which is in agreement with tentative assignment @xcite . while with jj44b and @xmath1 it is more than 200 kev higher .
the jj44b interaction predict gradual drop in 2@xmath37 energy from @xmath69ga to @xmath70ga .
this is reasonable trend because for @xmath70ga it is a g.s . which is in agreement with experimental data .
as we move from @xmath68ga onwards , the contribution in g.s .
wavefunction for @xmath71 configuration is start increasing .
there is also similarity between spectra of gallium and copper isotopes such as 2@xmath37 is a g.s . in case of @xmath72cu and @xmath73ga .
the leading proton configuration in @xmath72cu is @xmath74 which show an agreement with experimental moments .
[ t_be2 ] [ t_quad ]
as proposed in @xcite possibility of proton excitations across z=28 may play important role , by including the proton @xmath16 orbital in the model space .
we have calculated static quadrupole moments with effective charges @xmath75 and magnetic moments with @xmath76 in @xmath77 model space as shown in table [ t_quad ] .
the @xmath1 interaction predict correct sign of quadrupole moments for @xmath78ga and results are better than if we use only @xmath0 space @xcite .
the change in sign of quadrupole moments from @xmath35ga , i.e. @xmath2 onwards , demonstrate a changing of shell structure .
this is due to ground state of @xmath35ga have @xmath47 configuration ( @xmath22 28% ) and @xmath79ga have @xmath80 configuration .
these configurations reveal change of sign of quadrupole moments because filling of higher orbital started . thus below @xmath11 , a hole configuration ( @xmath47 )
has a positive quadrupole moment and above @xmath11 , a particle configuration ( @xmath80 ) has a negative quadrupole moment .
the @xmath1 interaction predict correct sign of magnetic moment for @xmath7ga while jun45 and jj44b gives negative sign .
this support ground state as 1/2@xmath37 for @xmath7ga .
we have also reported quadrupole moments for @xmath81ga isotopes with @xmath1 interaction .
for @xmath69ga the @xmath1 give reasonable value of quadrupole moment in comparison to jj44b @xcite while it predict too low value for @xmath68ga .
in the present work large scale shell model calculations have been performed for @xmath13ga isotopes in @xmath0 and @xmath17 model spaces with @xmath82ni and @xmath28ca core respectively . for the @xmath0 space ,
calculations have been performed with jun45 and jj44b interactions and for @xmath17 model space with @xmath1 interaction . in @xmath17 space
, we use a truncation by allowing maximum three particle excitations .
our calculated results predict high - spin sequences built on top of the 3/2@xmath37 , 5/2@xmath37 and 9/2@xmath36 levels in odd @xmath4ga .
the results of @xmath1 interaction are better than jun45 and jj44b in lighter isotopes . while for heavier isotopes , the results of jj44b interaction are quantitatively better than jun45 .
the results of level energies for heavier ga isotopes with @xmath1 interaction are more compressed .
the calculated @xmath43 values for @xmath7ga with jj44b interaction give reasonable agreement with experimental data .
we have also pointed out that the energy level systematics of ga isotopes are quite similar to that in cu isotopes .
there is a clear need for suitable interaction for neutron - rich nuclei in this region for better understanding of nuclear structure .
thanks are due to v.k.b .
kota , m. honma , j. diriken ( for @xmath43 values ) , r. sahu and i. mehrotra for useful discussions .
i would also like to thank f. nowacki for supplying @xmath1 interaction matrix elements .
99 i. stefanescu _ et al .
_ , phys . rev .
c * 79 * 064302 ( 2009 ) . | we have performed comprehensive set of shell model calculations for ga isotopes including high - spin states with three different effective interactions .
this work will add more information in the earlier work by cheal _
et al . _ for odd - even ga isotopes [ phys . rev .
lett .
104 , 252502 ( 2010 ) ] and man _ et al .
_ for odd - odd ga isotopes [ phys .
rev .
c 84 , 024303 ( 2011 ) ] , where only few excited states are studied in @xmath0 space . for lighter isotopes
@xmath1 interaction is better and for heavier isotopes jj44b is quantitatively better than jun45 .
these results show that limitation of existing interactions and calling for further improvements to predict nuclear structure properties of ga isotopes . |
there are two basic low energy features of strong interactions , which should be explained by qcd : confinement and spontaneous breaking of chiral symmetry .
while the latter can be more or less explained in terms of instantons @xcite , confinement is much less understood .
recently , however , much evidence has been accumulated for the idea that confinement is realized as a dual meissner effect @xcite , at least in the so - called abelian gauges @xcite . in these gauges
magnetic monopoles arise , which have to condense in order to form the dual superconducting ground state . from the conceptional point of view
it is unsatisfactory that spontaneous breaking of chiral symmetry and confinement is attributed to different types of field configurations .
this is because lattice calculations show that both phenomena occur at about the same energy scale .
in fact , lattice calculations indicate that the deconfinement phase transition and the restoration of chiral symmetry occur at about the same temperature . furthermore , for the spontaneous breaking of chiral symmetry the topological properties of the instantons
are crucial , giving rise to zero modes of the quarks , which in turn lead to a non - zero level density at zero virtuality .
the latter is sufficient to ensure a non - zero quark condensate , which is the order parameter of chiral symmetry breaking @xcite . on the other hand ,
magnetic monopoles are long - range fields , which should also be relevant for the global topological properties of gauge fields .
we can therefore expect an intimate relation between magnetic monopoles and the topology of gauge fields .
it is the aim of my talk to explain how the non - trivial topology of gauge fields is generated in abelian gauges by magnetic monopoles @xcite .
magnetic monopoles arise in the so - called abelian gauges , which fix the coset @xmath0 of the gauge group @xmath1 , but leave the abelian gauge invariance with respect to the cartan subgroup @xmath2 @xcite .
recent lattice calculations show that the dual meissner effect is equally well realized in all abelian gauges studied @xcite , while the abelian and monopole dominance is more pronounced in the so - called maximum abelian gauge @xcite . here , for simplicity , we will use the polyakov gauge , defined by a diagonalization of the polyakov loop @xmath3 where @xmath4 is the diagonal polyakov loop and the matrix @xmath5 is obviously defined only up to an abelian gauge transformation @xmath6 .
this gauge is equivalent to the condition @xmath7 where @xmath8 and @xmath9 denote the diagonal ( neutral with respect to @xmath2 ) and off - diagonal ( charged ) parts of the gauge field , respectively . when the polyakov loop ( [ x1 ] ) becomes a center element of the gauge group at some isolated point @xmath10 in @xmath11-space
there is a topological obstruction in the diagonalization and the induced gauge field @xmath12 arising from the gauge transformation @xmath13 which makes the polyakov loop diagonal develops a magnetic monopole @xcite .
let us illustrate this for the gauge group @xmath14 , where the cartan subgroup is @xmath15 and the polyakov loop can be parametrized as @xmath16 with @xmath17 being the pauli matrices . in this case the matrix @xmath5 can be chosen as @xcite @xmath18 where @xmath19 and @xmath20 denote the polar and azimuthal angles of the color unit vector @xmath21 , respectively , and the induced gauge field ( [ x4 ] ) acquires the form @xmath22 at those isolated points @xmath23 in space where @xmath24 , @xmath25 being an integer , and consequently the polyakov loop becomes a center element @xmath26 the abelian part of the induced gauge potential @xmath27 ( [ x6 ] ) has the form of a magnetic monopole . to be more precise the abelian part of the magnetic field
, @xmath28 develops an ordinary dirac monopole with a dirac string along the negative 3-axis . note that the @xmath29 gauge freedom in the definition of @xmath13 can be exploited to move the dirac string around .
the abelian component of the non - abelian part of the magnetic field @xmath30 , contains an oppositely directed monopole field , but without a dirac string . in the abelian component of the total magnetic field @xmath31 ^ 3 $ ]
the monopole field cancels and only the dirac string survives @xcite .
this is easily understood by noticing that the field @xmath27 ( [ x4 ] ) is a pure gauge and its field strengths @xmath32 $ ] should hence vanish unless there are singularities , which here occur on the dirac string .
an important property of the magnetic monopoles arising in the abelian gauges is that their magnetic charge @xmath33 & = & \frac{1}{4 \pi } \int_{\bar{s}_2 } d \vec{\sigma } \vec b^3\nonumber\nonumber\\ & = & - \frac{1 } { 8 \pi } \int_{s_2 } d \sigma_i \epsilon_{ijk } [ { \cal a}_i , { \cal a}_j]\label{x10}\end{aligned}\ ] ] is topologically quantized and given by the winding number @xmath34 \in \pi_2 ( su(2)/u(1))$ ] of the mapping @xmath35 . in the above equation @xmath36 is an infinitesimal @xmath37-sphere around the monopole position and @xmath38 arises from @xmath36 by leaving out the position of the dirac string .
since the magnetic monopoles are long - range fields , we expect that they are relevant for the topological properties of gauge fields .
as is well known , the gauge fields @xmath39 are topologically classified by the pontrjagin index @xmath40 in the gauge @xmath41 ( which is satisfied by the polyakov gauge ) and for temporally periodic spatial components of the gauge field @xmath42 , the pontrjagin index @xmath43 $ ] ( [ x11 ] ) equals the winding number of the polyakov loop @xmath44 @xcite : @xmath45 & = & \frac{1}{24 \pi^2 } \int d^3 x \epsilon_{ijk } \text{tr } l_i l_j l_k\\ ( l_i & = & \omega \partial_i \omega^\dagger)\nonumber.\end{aligned}\ ] ] for this winding number to be well defined , @xmath46 has to approach at spatial infinity @xmath47 a value independent of the direction @xmath48 , so that the @xmath11-space @xmath49 can be topologically compactified to the 3-sphere @xmath50 . without loss of generality
we can assume that @xmath51 approaches a center element @xmath52 { } ( -1)^{n_0}\label{x13}\end{aligned}\ ] ] with @xmath53 integer . in @xcite the following exact relation between the winding number @xmath54
$ ] and the magnetic monopole charges has been derived @xmath55 = \sum_i l_i m_i.\label{x14}\end{aligned}\ ] ] here the summation runs over all magnetic monopoles , @xmath56 being the monopole charges , and the integer @xmath57 is defined by the center element @xmath58 which the polyakov loop becomes at the monopole position @xmath23 ( cf .
eq . [ x3 ] ) , and by the boundary condition eq .
[ x13 ] . for an isolated monopole with a dirac string running to infinity the quantity @xmath59 represents the length of the dirac string in group space . to illustrate the above relation ( [ x14 ] )
consider the familiar instanton , for which the temporal component of the gauge field is given by @xmath60 with @xmath61 being the instanton size .
the polyakov loop for the instanton field acquires the form of a hedgehog @xmath62 in the center of the hedgehog there is a magnetic monopole @xmath63 , for which @xmath64 , and which carries a magnetic charge @xmath65 .
furthermore , the hedgehog , eq . [ x17 ] , satisfies the boundary condition eq .
[ x13 ] with @xmath66 . in this case
the length of the dirac string in group space becomes @xmath67 and the winding number becomes @xmath54= -1 $ ] , which is the correct result .
for the sake of illustration let us also consider a monopole - anti - monopole pair with opposite magnetic charges connected by a dirac string .
we denote the value of the polyakov loop at the monopole and anti - monopole positions by @xmath68 and @xmath69 , respectively . from eq .
[ x14 ] we then find for the winding number @xmath70 & = & ( n - n_0 ) m-(\bar n - n_0 ) m\nonumber\\ & = & ( n - \bar n ) m.\end{aligned}\ ] ] again the winding number of the polyakov loop field is given by the product of the magnetic charge @xmath71 times the length of the dirac string @xmath72 connecting monopole and anti - monopole .
the asymptotic value of the polyakov loop field at spatial infinity ( [ x13 ] ) drops out from this expression .
finally let us replace our spatial manifold @xmath49 by the compact space @xmath50 . in this case
the point at spatial infinity becomes part of the manifold , and there is an extra monopole due to our boundary condition at spatial infinity , eq . [ x13 ] , collecting all the dirac monopoles running to infinity . noticing that on a compact manifold the net charge has to vanish , @xmath73 , the monopole at infinity must carry the opposite net charge @xmath74 of all the remaining magnetic monopoles @xmath75 .
. [ x14 ] then simplifies to @xmath45 = \sum_{i=0 } n_i m_i,\label{x19}\end{aligned}\ ] ] where the summation over the monopoles now also includes the monopole at infinity @xmath76 .
[ x14 ] shows that in the polyakov gauge the topology of gauge fields is exclusively determined by the magnetic monopoles .
given the fact that all abelian gauges considered so far show the dual meissner effect for the yang mills vacuum
we conjecture that this result is not restricted to the polyakov gauge but applies to all abelian gauges .
the above considerations show that magnetic monopoles are entirely sufficient to generate a non - trivial topology and hence a non - zero topological susceptibility
. therefore , magnetic monopoles are also sufficient to trigger spontaneous breaking of chiral symmetry , which is usually attributed to instantons .
instantons in abelian gauges give rise to magnetic monopoles as we have seen above , but monopoles can exist in the absence of instantons .
the above considerations show that in the abelian gauges both confinement and spontaneous breaking of chiral symmetry can be generated by the same field configurations : magnetic monopoles .
discussions with m. , k. langfeld , m. quandt , a. schfke and o. are greatly acknowledged .
99 h. leutwyler and a. smilga , phys . rev . *
d 46 * ( 1992 ) 5607 and references therein .
s. bali , ch .
schlichter and k. schilling , hep - lat/9802005 and references therein .
g. t hooft , nucl . phys .
* b 190 * ( 1981 ) 455 .
h. reinhardt , nucl .
* b 503 * ( 1997 ) 505 . a. di giacomo , this meeting .
k. langfeld , m. quandt , h. reinhardt , hep - th/9619213 .
a. s. kronfeld , g. schierholz and u .- j .
wiese , nucl .
* b 293 * ( 1987 ) 461 . | lattice calculations performed in abelian gauges give strong evidence that confinement is realized as a dual meissner effect , implying that the yang mills vacuum consists of a condensate of magnetic monopoles .
we show in polyakov gauge how the pontrjagin index of the gauge field is related to the magnetic monopole charges . |
in this article we are interested in what we will call _ universal mapping class groups _ of genus @xmath0 ; these are groups that contain the mapping class group of every compact orientable surface of genus @xmath3 with non - empty boundary .
universal mapping class groups appear naturally in various contexts . as a first example
, one can consider a set @xmath10 of compact surfaces with boundary , equipped with the partial order given by surface inclusion , and take the direct limit of the corresponding mapping class groups . by construction , this direct limit is an infinitely - generated group which contains the mapping class group @xmath11 for every @xmath12 .
this is the approach taken by harer @xcite , and subsequently by many others , in order to study the so - called _ stable homology _ of the mapping class group ; compare with the discussion around theorem [ thm : homology ] below .
a more general approach is to consider the mapping class group of an appropriate _ infinite - type _ surface , i.e. with infinitely - generated fundamental group . in this direction , consider the surface @xmath2 that results from removing a cantor set from a closed orientable surface of genus @xmath13 .
then its mapping class group @xmath9 , i.e. the group of isotopy classes of self - homeomorphisms of @xmath2 , contains the mapping class group of every compact surface of genus @xmath3 with non - empty boundary .
the groups @xmath9 were recently ( re-)brought into the picture by calegari @xcite , who remarked that every group that acts freely by homeomorphisms on a closed surface of genus @xmath3 admits a natural homomorphism to @xmath9 .
observe , however , that @xmath9 is certainly enormous : for instance , it surjects onto the homeomorphism group of a cantor set ( see e.g. below ) , which is itself uncoutable . before continuing we
remark that , quite surprisingly , if one were to define @xmath9 in terms of isotopy classes of _ ambient diffeomorphisms _ instead , then the resulting group turns out to be discrete , as proved recently in @xcite .
in any of the approaches described above , we see that the resulting universal mapping class group is ( at least ) infinitely - generated , and thus not that attractive from the point of view of geometric group theory .
this obstacle was first overcome by @xcite in the case of genus zero .
more concretely , in that paper it is proved that the subgroup @xmath14 of _ asymptotically rigid _
mapping classes is finitely presented , and still contains the mapping class group of every compact surface of genus zero .
roughly speaking , the elements of @xmath15 are mappping classes that preserve some kind of _ rigid structure at infinity _ , see section [ sec : defs ] for a formal definition .
subsequently in @xcite , the same authors studied a similar asymptotically rigid mapping class group @xmath16 , this time for a surface @xmath17 of infinite genus with a cantor space of ends .
they proved that @xmath16 is finitely generated , and contains the mapping class group of _ every _ compact surface with non - empty boundary .
moreover , the homology of @xmath16 coincides with the homology of the stable mapping class group .
the main objective of this article is to extend the definition from @xcite of an asymptotically rigid mapping class group ( or asymptotic mapping class group , for short ) @xmath4 , to every finite genus @xmath18 . by construction
, the group @xmath4 will be a subgroup of @xmath9 , and will contain the mapping class group of every compact surface of genus @xmath3 with non - empty boundary .
we will then prove that the groups @xmath4 so obtained behave much like their finite - type counterparts .
first of all , we will prove : @xmath4 is finitely presented , for every @xmath19 .
[ thm : fg ] next , along the lines of @xcite , we will observe that @xmath4 is also universal in a homological sense . more concretely ,
let @xmath20 denote the connected orientable surface of genus @xmath3 with @xmath21 boundary components .
a celebrated result of harer @xcite asserts that , for a fixed genus @xmath22 , the @xmath5-th homology group of the _ pure _ mapping class group @xmath23 is independent of @xmath21 , provided that @xmath24 ; for this reason , this homology group is called the @xmath5-th _ stable homology group _ of the mapping class group of genus @xmath3 . using a translation of the proof of (
* theorem 3.1 ) to our setting , we will show : let @xmath25 .
for every @xmath26 , the @xmath5-th homology group @xmath27 is isomorphic to the @xmath5-th stable homology group of the ( pure ) mapping class group of genus @xmath3 .
[ thm : homology ] a result of powell @xcite asserts that @xmath28 is _ perfect _ , i.e. has trivial abelianization . as a direct consequence of theorem [ thm : homology ] , we obtain the following analogue for asymptotically rigid mapping class groups : @xmath4 is perfect for every @xmath22 .
combining our definition of the groups @xmath4 with the corresponding ones of @xmath15 and @xmath16 , from @xcite and @xcite
respectively , we now have a family @xmath29 of finitely - generated groups ( in fact , finitely - presented if @xmath30 ) such that @xmath4 contains the mapping class group of every compact surface of genus @xmath3 with non - empty boundary .
next , we proceed to study some properties of this family , in analogy with known and/or expected theorems about finite - type mapping class groups . in the light of the results above ,
a first obvious question is whether the groups we have defined are pairwise distinct .
this is indeed the case for mapping class groups of finite - type surfaces , where two of them are isomorphic if and only if the underlying surfaces are homeomorphic .
this follows , for instance , as a combination of harer s calculation of the virtual cohomological dimension of the mapping class group @xcite , plus the computation of the maximal rank of its abelian subgroups @xcite .
observe that , in the case of the groups @xmath4 , both of these quantities are infinite . in spite of this
, we will prove that the isomorphism type of @xmath4 is completely determined by @xmath3 .
in fact , this will follow as a consequence of a stronger result , although we need a definition before stating it .
say that a homomorphism @xmath31 between two groups is _ @xmath32-injective _ if for every subgroup @xmath33 with @xmath34 , we have that @xmath35 is injective .
we will prove : let @xmath36 . if @xmath37 , then there are no @xmath32-injective maps from @xmath38 to @xmath4 .
[ thm : ainj ] since @xmath4 contains a subgroup isomorphic to @xmath39 ( see lemma [ lem : a - inj ] ) for all @xmath3 , and isomorphisms are @xmath32-injective , we obtain : let @xmath36 .
then @xmath40 if and only if @xmath41 .
[ cor : iso ] a related question is to determine the automorphism group of @xmath4 .
recall that a celebrated theorem of ivanov @xcite asserts that the mapping class group of a ( sufficiently complicated ) finite - type surface is _ rigid _ , that is , every automorphism is induced by a homeomorphism of the underlying surface .
this result has recently been extended to the infinite - type setting by patel - vlamis @xcite , who proved that the pure mapping class group of @xmath2 is rigid . along similar lines
, our next result asserts that @xmath4 is also rigid .
more concretely , denoting by @xmath42 the automorphism group of @xmath43 , we will show : for every @xmath30 , @xmath44 coincides with the normalizer @xmath45 of @xmath4 within @xmath9 .
[ thm : autos ] an immediate consequence of ivanov s theorem in the finite - type case is that the outer automorphism group of @xmath46 is finite ; in fact , it is trivial for all but finitely many surfaces .
however , in corollary [ cor : out ] we will see that this is no longer true for asymptotically rigid mapping class groups . more concretely ,
denote by @xmath47 the outer automorphism group of a group @xmath43 , that is , the group of conjugacy classes of automorphisms of @xmath43 .
we will prove : for every @xmath30 , @xmath48 is an infinite group .
[ cor : out ] building up on work of ivanov @xcite and kaimanovich - masur @xcite , farb - masur @xcite proved that any homomorphism from a higher - rank lattice to a finite - type mapping class group has finite image ( see also @xcite for different proofs of this result ) . in short ,
our next theorem is an analogue of farb - masur s theorem in the context of asymptotically rigid mapping class groups : let @xmath49 be a lattice in a semisimple lie group @xmath43 of real rank at least two , where @xmath43 has no compact factors isogenous to @xmath50 or @xmath51 . for every @xmath52 ,
any homomorphism @xmath53 has finite image .
[ thm : lattice ] a compactly generated group has _
kazhdan s property ( t ) _ if every unitary representation that has almost invariant vectors also has an invariant vector . since we will not need any further details about property ( t ) , we simply refer the reader to the book @xcite for details .
it is expected @xcite that mapping class groups of finite - type surfaces do not have property ( t ) . using similar arguments to the ones in the proof of theorem [ thm : lattice ]
, we will observe : let @xmath54 .
then @xmath4 does not have property ( t ) .
[ thm : t ] a well - known outstanding question about finite - type mapping class groups asks whether they are _ linear _ , that is , whether they admit an injective homomorphism into @xmath55 for some @xmath56 . the only known result in this direction is a theorem of bigelow - budney @xcite , who proved that the mapping class group of the closed surface of genus two is linear , by exploiting its relation with braid groups . in sharp contrast , in proposition [ split ] we will see that @xmath4 contains an isomorphic copy of thompson s group @xmath57 , for all @xmath30 . combining this with theorem [ thm : fg ] , we have : if @xmath30 then @xmath4 is not linear .
[ thm : nonlinear ] finally , in the appendix we will study the group of _ half - twists _
@xmath8 of @xmath2 , where @xmath30 .
the group @xmath8 is a close relative of @xmath4 , whose elements are isotopy classes of _ asymptotically quasi - rigid _ homeomorphisms of @xmath2 ; see section [ sec : rigid ] . as will become evident from the definitions , @xmath58 .
we remark that the group @xmath59 first appeared ( under the name @xmath60 ) @xcite , where it is proved that it is finitely generated , with respect to an explicit generating set . using similar arguments as the ones employed with @xmath4
, we will prove that @xmath8 is in fact finitely presented for all @xmath3 , and that its automorphism group is again isomorphic to its normalizer within @xmath9 .
however , the main interest of the group @xmath8 resides in our next result , which asserts that @xmath8 serves as a finitely - presented approximation of @xmath9 ; recall that this latter group is uncountable . more concretely , endowing @xmath9 with the quotient topology coming from the compact - open topology in @xmath61 , we will prove : for every @xmath30 , the group @xmath8 is dense in @xmath9 .
[ thm : limit ] in a related direction , a consequence of a recent result of patel - vlamis @xcite is that @xmath9 is topologically generated by dehn twists .
* plan of the paper . * in section [ sec : defs ] we review some basic definitions on surfaces and their mapping class groups . in section [ sec : rigid ] we will define the notion of a rigid structure on the surfaces @xmath2 .
this will be used in section [ sec : as - mcg ] to define asymptotically rigid mapping class groups . in section [ sec :
v ] we will explore the relation of asymptotically rigid mapping class groups and thompson s group @xmath62 by means of a short exact sequence , which will be a key ingredient in the proof of many of our main results .
we then proceed to prove the results mentioned in the introduction . in this direction ,
theorem [ thm : fg ] will be proved in section [ sec : prooft1 ] .
section [ sec : homology ] will deal with the proof of theorem [ thm : homology ] . in section [ sec : injective ] we will prove theorem [ thm : ainj ] , while section [ sec : proofaut ] is concerned with the proof of theorem [ thm : autos ] . in section [ sec : lattices ] we will establish theorems [ thm : lattice ] and [ thm : t ] .
finally , in the appendix we will introduce the group of half - twists and prove it is dense in the mapping class group .
* acknowledgements .
* the authors thank y. antoln , j. bavard , j. hernndez , c. kapoudjian , t. koberda , c. martnez - prez , h. parlier , j. souto , and n. vlamis for conversations .
we are indebted to juan souto for suggesting the proof of lemma [ lem : twists2twists ] .
finally , we thank the authors of @xcite for sharing their preprint with us .
in this section we recall some of the basic definitions about surfaces and their mapping class groups .
let @xmath63 be a connected orientable surface of negative euler characteristic .
if the fundamental group of @xmath63 is finitely generated , we will say that @xmath63 has _ finite type _ ; otherwise , we say that it has _
infinite - type_. by a _ curve _ on @xmath63 we mean the isotopy class of a simple closed curve on @xmath63 which does not bound a disk , a once - punctured disk , or an annulus whose other boundary curve is a boundary component of @xmath63 .
we say that two curves are _ disjoint _ if they have disjoint representatives on @xmath63 ; otherwise we say that they intersect .
given curves @xmath64 , we define their _ intersection number _ @xmath65 as the minimal number of points of intersection between representatives .
a _ multicurve _ on @xmath63 is a set of pairwise disjoint curves on @xmath63 .
we say that a multicurve @xmath66 is _ locally finite _ if any compact subsurface of @xmath63 intersects only finitely many elements of @xmath66 .
a _ pants decomposition _ of @xmath63 is a locally - finite multicurve that is maximal with respect to inclusion ; as such , its complement on @xmath63 is a disjoint union of 3-holed spheres , or
_ pairs of pants_. finally , an _ arc _ on @xmath63 is a non - trivial isotopy class of properly embedded arcs on @xmath63 .
it is well - known that the homeomorphism type of a finite - type surface is completely determined by its genus and number of punctures and/or boundary components . in a similar fashion
, there is also a classification of infinite - type surfaces up to homeomorphism @xcite , this time in terms of its genus and the topological type of its _ space of ends_. we now briefly recall these definitions and results , referring the reader to @xcite for a thorough discussion .
let @xmath63 be a connected orientable surface of infinite type . by an _ ending sequence _ on @xmath63 we mean
a collection @xmath67 of connected open subsets of @xmath63 , such that @xmath68 is not relatively compact , @xmath69 is compact , and every relatively compact subset of @xmath63 is disjoint from all but finitely many @xmath68 .
we declare two ending sequences to be _ equivalent _ if every element of one is contained in some element of the other .
an _ end _ of @xmath63 is then an equivalence class of ending sequences .
we will write @xmath70 for the set of ends of @xmath63 , which becomes a topological space after specifying the following basis for a topology : given a subset @xmath71 with compact boundary , let @xmath72 be the set of all ends of @xmath63 that have a representative exiting sequence that is eventually contained in @xmath73 .
the following theorem is a special case of theorem 1.5 in @xcite : let @xmath63 be a connected orientable surface of infinite topological type . then @xmath70 is totally disconnected , separable , and compact ; in particular , it is a subset of a cantor set . [
thm : ends ] we now recall the classification of infinite - type surfaces up to homeomorphism @xcite . before doing so
, we need a definition .
let @xmath63 be a connected orientable surface of infinite type .
planar end _ of @xmath63 is a end defined by an ending sequence whose elements are eventually planar .
otherwise , the end is said to be _ non - planar_. write @xmath74 and @xmath75 , respectively , for the subspaces of planar and non - planar ends of @xmath63 .
clearly @xmath76 . in @xcite , richards proved : let @xmath63 and @xmath77 be two connected orientable surfaces .
then @xmath63 and @xmath77 are homeomorphic if and only if they have the same genus , and @xmath78 is homeomorphic to @xmath79 as nested topological spaces .
that is , there exist a homeomorphism @xmath80 whose restriction to @xmath75 defines a homeomorphism between @xmath75 and @xmath81 .
[ thm : class ] in what follows we will concentrate solely on connected orientable surfaces of infinite topological type and _ without isolated ends_. note that , in the light of theorems [ thm : ends ] and [ thm : class ] , such a surface is homeomorpic to either a closed surface of finite genus @xmath13 with a cantor set removed , or to a surface of infinite genus whose space of ends is homeomorphic to a cantor set .
we remark that the latter is sometimes known as the _ blooming cantor tree_. * notation .
* we will write @xmath2 for the surface obtained by removing a cantor set from a closed surface of genus @xmath3 , and @xmath17 for the surface of infinite genus with no isolated ends and whose space of ends is homeomorphic to a cantor set . in the finite - type case
, we will denote by @xmath20 the connected orientable surface of genus @xmath3 with @xmath21 boundary components .
let @xmath63 be a connected orientable surface , possibly of infinite type .
the _ mapping class group _
@xmath82 is the group of isotopy classes of self - homeomorphisms of @xmath63 ; if @xmath63 has boundary , we will assume that it is parametrized , and require that homeomorphisms and isotopies respect this parametrization .
in particular , an element of @xmath82 may induce a non - trivial permutation on the set of boundary components of @xmath63 . before
continuing we observe that , while @xmath46 is finitely presented ( and hence discrete ) the group @xmath9 is uncountable ; indeed , it surjects onto the homeomorphism group of a cantor set , see the exact sequence below .
we record the following immediate observation for further use : let @xmath52 .
then @xmath83 is a subgroup of @xmath9 , for every @xmath84 and every @xmath85 ( in fact , for @xmath86 if @xmath87 ) . in what follows
, we will need to make use of the following subgroups of @xmath82 .
the _ pure mapping class group _
@xmath88 is the subgroup of @xmath82 whose elements induce the trivial permutation on the space of ends and fix each boundary component pointwise . if @xmath89 , the _ compactly supported _
( pure ) mapping class group @xmath90 is the subgroup of @xmath9 whose elements are the identity outside a compact subsurface of @xmath2 , called a _ support _ ; note that such a mapping class may be naturally regarded as an element of the mapping class group of a finite - type surface . finally , observe that both @xmath91 and @xmath90 are infinitely - generated groups .
in this section , we briefly review the definition of rigid structure for @xmath92 and @xmath17 , introduced in @xcite and @xcite respectively , and generalize this notion to arbitrary @xmath3 .
we recall the definition of rigid structure for @xmath92 from @xcite ; the reader should keep figure [ fig : genus0 ] in mind .
a _ rigid structure _ on @xmath92 is a triple @xmath93 , where : * @xmath94 is a pants decomposition of @xmath92 , called the _ pants decomposition underlying the rigid structure _ , * @xmath32 is a set of arcs on @xmath92 with the following properties : 1 .
each arc joins two distinct ends @xmath92 , 2 . for every pair of pants @xmath95 ,
there are exactly three elements of @xmath32 intersecting @xmath96 essentially , each of them connecting two different boundary curves of @xmath96 , * @xmath97 is one of the two connected components of @xmath98 , called the _ visible side _ of @xmath92 .
.,width=288,height=192 ] observe that , up to the action of @xmath99 , there is only one rigid structure on @xmath92 .
it will be useful later to extend the concept of a rigid structure to the surface @xmath100 obtained from @xmath92 by removing a finite collection of @xmath101 points .
we then define a rigid structure on @xmath100 as a rigid structure on the surface obtained from @xmath100 by filling every isolated puncture ( which is homeomorphic to @xmath92 ) , subject to the condition that every isolated puncture of @xmath100 is contained in the same connected component of @xmath102 , where @xmath94 is the pants decomposition underlying the rigid structure of @xmath92 .
consider the case of @xmath103 .
a _ rigid structure _ on @xmath2 consists of a curve @xmath104 that cuts off a disk containing every puncture of @xmath2 , together with a rigid structure for the planar component of @xmath105 , namely the one homeomorphic to @xmath106 .
we now recall the definition from @xcite of rigid structure for @xmath17 . to this end
, it is useful to realize @xmath17 as follows ; the reader should keep figure [ fig : s - infty ] for this discussion .
let @xmath107 be the infinite trivalent tree , and for simplicity consider it as embedded in @xmath108 inside @xmath109 .
now , glue an _ edge - loop _ ( that is , an edge with a loop at one of its vertices ) to the midpoint of every edge of @xmath107 , such that the edge of the edge - loop is vertical .
denote by @xmath110 the graph so obtained , and observe that the boundary of a thickening of @xmath110 is homeomorphic to @xmath17 .
moreover , the graph @xmath110 is embedded inside @xmath17 as a cross - section of the obvious fiber projection .
.,width=384,height=240 ] this realization of the surface equips @xmath17 with a preferred pants decomposition , which we denote @xmath94 ; see figure [ fig : s - infty ] .
the _ rigid structure _ for @xmath17 is the pair @xmath111 , where we regard @xmath110 as embedded in @xmath17 .
again , we call @xmath94 the _ pants decomposition underlying the rigid structure_. before finishing this subsection , it will be useful to distinguish the following two types of pairs of pants in @xmath112 : * a _ vertex pair of pants _ is a pair of pants in @xmath113 such that each of the boundary curves separates @xmath17 into two connected components of infinite genus .
intuitively , such a pair of pants corresponds to a vertex of @xmath107 . * an _ edge pair of pants _ is a pair of pants in @xmath112 whose boundary curves are separating , and one of them cuts bounds a one - holed torus in @xmath17 .
in this section we will introduce the asymptotically rigid mapping class groups ( or asymptotic mapping class groups , for short ) @xmath4 for arbitrary @xmath52 . informally speaking , the elements of @xmath4 are mapping classes that asymptotically preserve the rigid structure defined in the previous section .
as mentioned above , we remark that the cases @xmath114 first were previously introduced in @xcite ; however , we also include them in the discussion here in order to make the discussion as self - contained as possible . as was the case in the previous section , the situation is slightly different if @xmath3 is finite or infinite , and hence we will discuss these cases separately .
let @xmath18 , and suppose @xmath2 has been endowed with the rigid structure defined in the previous section .
as usual , write @xmath94 for the pants decomposition underlying this rigid structure .
we say that a compact subsurface @xmath115 is _ @xmath94-suited _ if @xmath116 .
we now extend the definition of _ asymptotically rigid _
homeomorphism from @xcite and @xcite to arbitrary finite @xmath3 : we say that a homeomorphism @xmath117 is _ asymptotically rigid _ if there exists a @xmath94suited subsurface @xmath115 with @xmath118 also @xmath94suited , and such that the restriction homeomorphism @xmath119 sends : * @xmath120 to @xmath121 , * @xmath122 to @xmath123 , * the visible side of @xmath124 to the visible side of @xmath125 .
if we drop the last requirement we say that the homomeorphism @xmath126 is _ asymptotically quasi - rigid _
( this generalization will be heavily used in the appendix ) .
[ def : asrig1 ] * notation . * given an asymptotically rigid ( resp .
quasi - rigid ) homeomorphism @xmath126 as above , we will refer to the subsurface @xmath127 in the definition above as a _ defining subsurface _
_ generalized defining subsurface _ ) for @xmath126 .
we now review the definition of an asymptotically rigid homeomorphism of the surface @xmath17 given in @xcite .
recall from definition [ def : asrig2 ] that in this case the rigid structure consists of a pants decomposition @xmath94 and a graph @xmath110 embedded in @xmath17 .
we say that a compact subsurface @xmath127 of @xmath17 is _ @xmath94suited _ if the following holds : 1 .
every component of @xmath128 is a boundary curve of a vertex pair of pants in @xmath113 , 2 .
if a boundary curve of an edge pair of pants @xmath129 is contained in @xmath128 , then @xmath130 . as for asymptotically rigid homeomorphisms
, we have : we say that a homeomorphim @xmath131 is _ asymptotically rigid _ if : * there exists a @xmath94suited subsurface @xmath132 such that @xmath118 is also @xmath94suited . *
the restriction homeomorphism @xmath133 sends : * * @xmath134 to @xmath135 ; * * @xmath136 to @xmath137 [ def : asrig2 ] * notation .
* as in the finite - genus case , we will refer to the surface @xmath127 in the definition above as a _ defining subsurface _ for the homeomorphism @xmath126 .
we are finally in a position to define the asymptotic mapping class group : let @xmath138 .
the _ asymptotic mapping class group _ @xmath4 ( resp .
@xmath139 , @xmath16 ) is the group of isotopy classes of asymptotically rigid homeomorphisms of @xmath2 ( resp .
@xmath100 , @xmath17 ) .
recall that @xmath90 is the subgroup of @xmath9 whose elements have compact support and induce the identity map on the space of ends of @xmath2 .
observe that we have a natural inclusion @xmath140 we now give an alternate description of the group @xmath90 which will be useful in the sequel . to this end , denote again by @xmath94 the pants decomposition underlying the rigid structure on @xmath2 .
consider the set @xmath141 which has the structure of a partially ordered set with respect to subsurface inclusion .
then , the following is easily verified : for every @xmath52 , we have @xmath142 where @xmath63 varies in the family of @xmath94-suited subsurfaces of @xmath2 .
moreover , if @xmath30 then @xmath63 may be assumed to have genus @xmath3 .
[ lem : limit ]
as it turns out , asymptotic mapping class groups are strongly related to thompson s groups ; as a matter of fact , this relation will be behind the proof of most of our main results .
recall that _
thompson s group _
@xmath62 is the group of right - continous bijections of the unit circle that map the set of dyadic rationals to itself , are differentiable except at finitely many points , and on every interval of differentiability they are affine maps whose derivatives are powers of 2 .
we refer the reader to the standard reference @xcite for a thorough discussion on the different thompson s groups . extending results of @xcite
, we will prove : for every @xmath52 , there is a short exact sequence @xmath143 [ prop : sesb ] when @xmath114 , proposition [ prop : sesb ] is shown in @xcite and @xcite , respectively .
furthermore , in those papers it is proved that the sequence does not split in these cases , although it does split over thompson s group @xmath144 when @xmath145 ; compare with proposition [ split ] . in order to prove proposition [ prop : sesb ] , it will be useful to work with a different incarnation of the group @xmath62 .
namely , every element of @xmath62 is encoded by a _ symbol _
@xmath146 , where @xmath144 and @xmath147 are finite , _ reduced _ rooted trees with the same number of leaves , and @xmath148 is a bijection between the set of leaves of @xmath144 and @xmath147 ; see @xcite for an expanded definition ( although this is essentially contained in the proof of proposition [ prop : sesb ] ) , as well as a proof of the equivalence of the two definitions of @xmath62 .
armed with this equivalent definition of thompson s group @xmath62 , we adapt the arguments of ( * ? ? ?
* section 2 ) in order to prove proposition [ prop : sesb ] as mentioned above , the cases @xmath145 and @xmath149 are covered in @xcite and @xcite , respectively .
thus , we assume that @xmath150 .
fix a rigid structure on @xmath2 , and let @xmath151 be the separating curve used to define the rigid structure , as described in [ subsec : finitegenus ] .
write @xmath94 for the pants decomposition underlying the rigid structure , which recall is a pants decomposition of the planar component of @xmath105
. the dual tree to this pants decomposition is naturally a rooted 3-valent tree @xmath107 , where the root corresponds to the unique pair of pants of @xmath152 having @xmath151 as boundary curve .
define an _
admissible subtree _ of @xmath107 as a subtree @xmath144 whose vertices are either 1-valent or 3-valent . as usual , a 1-valent vertex of @xmath144 is called a _
leaf_. the set of leaves of @xmath144 is denoted @xmath153 , and the cardinality of @xmath153 is called the _ level _ of @xmath144 . given a leaf @xmath154 of @xmath144 , we denote by @xmath155 the admissible tree obtained from @xmath144 by adding to @xmath144 the two vertices of @xmath156 adjacent to @xmath154 .
we remark that using the inclusion of @xmath144 and @xmath155 into @xmath107 , one may distinguish the two vertices of @xmath157 as the _ left _ and _ right _ vertices of @xmath158 .
we will denote these as @xmath159 and @xmath160 , respectively . as above
we define a _ symbol _ as a triple @xmath146 , where @xmath144 and @xmath147 are admissible subtrees of equal level , and @xmath161 is a bijection . given a symbol @xmath162 and @xmath163 , we say @xmath146 and @xmath164 are _ leaf - equivalent _ , where @xmath165 is the unique extension of @xmath148 to @xmath155 that preserves left and right .
let @xmath166 be the equivalence relation on the set of symbols generated by leaf - equivalence , and @xmath43 be the set of equivalence classes of symbols .
the set @xmath43 turns into a group as follows .
given two elements of @xmath43 , we first represent them as symbols @xmath146 and @xmath167 and set @xmath168 the identity element is the class of @xmath169 , where @xmath144 is any admissible subtree .
it is easy to see that the product operation in @xmath43 is well - defined and that , with respect to this operation , @xmath43 is isomorphic to thompson s group @xmath62 .
after this considerations , we define a homomorphism @xmath170 as follows .
let @xmath171 and consider a defining subsurface @xmath127 for @xmath126 .
up to enlarging @xmath127 , we may assume that @xmath127 has genus @xmath3 .
we associate to @xmath126 the symbol @xmath172 , where @xmath144 ( resp .
@xmath147 ) is the minimal admissible subtree of @xmath173 contained in @xmath127 ( resp .
@xmath118 ) , and @xmath148 is the bijection between the sets of leaves of @xmath144 and @xmath147 , respectively , induced by the permutation between the sets of boundary components of @xmath127 and @xmath118 given by @xmath126 .
now , one easily checks that this correspondence gives rise to a well - defined homomorphism @xmath170 , which is clearly surjective .
we claim that the kernel of this homomorphism is exactly @xmath90 . indeed , suppose @xmath171 fixes every end of @xmath2 , and consider a non - empty defining subsurface @xmath127 for @xmath126 . since @xmath126 does not permute ends we may assume , up to enlarging @xmath127 , that every component of @xmath174 is mapped to itself .
therefore , after further enlarging @xmath127 if necessary , we deduce that @xmath126 is the identity outside @xmath127 , and hence has compact support , as desired .
when @xmath175 the short exact sequence is a manifestation of a more general phenomenon , which we now explain .
consider the surface @xmath2 ; as mentioned above , the space of ends @xmath176 is homeomorphic to a cantor set , which we will denote @xmath177 in order to relax notation .
observe that any homeomorphism of @xmath2 induces a homeomorphism of @xmath177 , and thus we have a homomorphism @xmath178 now , every cantor set on the plane is _ tame _ , meaning that , up to homeomorphism , it is contained in some line ; in particular , it is homeomorphic to the standard triadic cantor set . a theorem of scpin ( see e.g. ( * ? ? ?
* theorem 1 ) ) states that any homeomorphism of the standard triadic cantor set @xmath179\subset { \mathbb{r}}^2 $ ] extends to a homeomorphism of @xmath180 ; moreover , this homeomorphism can be assumed to be identity outside a large enough ball .
these considerations imply the following lemma : the natural homomorphism @xmath181 is surjective .
[ lem : cantorsurj ] moreover , @xmath182 is a simple group @xcite , and thus the connected component of the identity in @xmath182 is trivial .
therefore , the homomorphism @xmath183 descends to a surjective homomorphism @xmath184 using a slight abuse of notation .
observe that the kernel of this homeomorphism is exactly the pure mapping class group @xmath91 , and thus we have a short exact sequence @xmath185 if we replace @xmath9 by the asymptotic mapping class group @xmath4 in equation , we obtain : @xmath186 where @xmath187 denotes the image of @xmath4 in @xmath182 under the homomorphism @xmath183 of .
we claim that we have already encountered the groups to the left and right of @xmath4 in equation , under other guises .
first , by the same argument that we used to determine the kernel of the homomorphism @xmath170 , we have : for every genus @xmath3 , we have @xmath188 .
[ lem : techcompact ] at this point , the combination of equations and , plus the first isomorphism theorem , together imply : @xmath187 is isomorphic to thompson s group @xmath62 .
in this section we prove that asymptotic mapping class groups of finite positive genus are finitely presented .
as mentioned in the introduction , the case of genus zero was originally settled in @xcite , and in fact will be a key ingredient in our arguments : the group @xmath15 is finitely presented .
[ thm : fk ] it will be useful for us to give a brief description of the arguments used in @xcite for proving theorem [ thm : fk ] .
the first ingredient , which will also play a central rle here , is the following well - known result of brown @xcite : let @xmath43 be a group acting on a connected and simply - connected cw - complex @xmath189 by permuting its cells .
suppose that : 1 .
the stabilizer of each vertex of @xmath189 is finitely presented ; 2 .
the stabilizer of every edge of @xmath189 is finitely generated ; 3 .
@xmath190 is cocompact
. then @xmath43 is finitely presented .
[ thm : brown ] in @xcite , funar - kapoudjian applied theorem [ thm : brown ] to the action of @xmath15 on a modification of the so - called _ pants complex _
@xcite of @xmath92 .
more concretely , they first consider a graph @xmath191 whose vertices are ( isotopy classes of ) pants decompositions of @xmath92 which differ from the pants decomposition underlying the rigid structure in a finite number of curves .
an edge of @xmath192 is given by two pants decompositions that are related by an _
elementary move _ , meaning that they differ in exactly two curves , which intersect exactly once ( resp .
twice ) if their union fills a one - holed torus ( resp . a four - holed sphere ) ; see @xcite for details
. using the same proof of the main result of hatcher - thurston @xcite , one deduces that the graph @xmath192 is connected , and that it becomes a simply - connected 2-complex after gluing a 2-cell to every _ triangle _ , _ square _ , and _ pentagon _ of @xmath192 .
however , it turns out that the action of @xmath15 on @xmath193 is not cocompact , as there are infinitely many @xmath15-orbits of squares ( * ? ? ?
* proposition 5.4 ) ; in particular , theorem [ thm : brown ] can not be applied to this situation . in order to overcome this , funar - kapoudjian construct a modification of @xmath192 , called the _ reduced _ pants complex , by considering only two combinatorial types of squares .
they then prove that @xmath15 acts satistying all the hypothesis of theorem [ thm : brown ] , as desired .
in fact , a minor variation of the arguments in @xcite serves to prove the following strengthening of theorem [ thm : fk ] , which will be crucial for us : for every @xmath194 , the group @xmath139 is finitely presented .
[ prop : extfk2 ] the definition of the reduced pants complex from @xcite makes sense also for @xmath100 , with arbitrary @xmath195 . by the same arguments as in @xcite , this
complex is connected and simply - connected and , using the same reasoning as in @xcite , @xmath139 acts cocompactly on it .
the stabilizer of every cell is an extension of a finite permutation group by a finitely generated free abelian group , and in particular finitely presented . at this point ,
the result follows from theorem [ thm : brown ] .
we now turn to the proof of theorem [ thm : fg ] , whose statement we now recall : theorem [ thm : fg ] for every @xmath19 , the group @xmath4 is finitely presented . in the light of theorem [ thm : fk ] , it suffices to show the result for @xmath3 positive . in order to do so
, we will also use brown s theorem [ thm : brown ] , this time using the action of @xmath4 on the so - called _ cut - system complex _
@xmath196 of @xmath2 .
we remark that cut - system complexes of finite - type surfaces were in fact used by hatcher - thurston @xcite and wajnryb @xcite in order to compute finite presentations of mapping class groups .
we now introduce the complex @xmath196 . for concreteness
, we choose to write the definitions for the surface @xmath2 , although they make sense for an arbitrary connected orientable surface , of finite or infinite type .
a _ cut system _ of @xmath2 is a set of @xmath3 curves whose union does not separate @xmath2 ; in particular , each of the curves is non - separating .
let @xmath196 be the simplicial graph whose vertices are cut systems on @xmath2 , and where two cut systems are adjacent in @xmath196 if they differ in exactly two curves , and these intersect exactly once . moreover ,
similar to the case of the pants complex , @xmath196 will become a simply - connected 2-complex after gluing a 2-cell to certain circuits in @xmath196 . before explaining this
, we need to borrow some definitions from @xcite .
a _ triangle _ in @xmath196 consists of three pairwise - adjacent vertices of @xmath196 ; geometrically , a triangle corresponds to three curves that pairwise intersect exaclt once .
square _ in @xmath196 is a closed path with four vertices @xmath197 such that @xmath198 and @xmath199 are adjacent in @xmath196 if and only if @xmath200 ; geometrically , a square corresponds to two elementary moves that occur in two disjoint one - holed tori . finally , a _ pentagon _ in @xmath196 consists of five vertices @xmath201 such that @xmath198 and @xmath199 are adjacent in @xmath196 if and only if @xmath202 ; geometrically
, a pentagon is determined by five curves @xmath203 on @xmath2 such that @xmath204 and @xmath205 , counting indices modulo 5 .
armed with these definitions , we construct a 2-complex by gluing a 2-cell to every triangle , square , and pentagon of @xmath196 . by a slight abuse of notation , we will denote the resulting complex by @xmath196 also .
the following result is due to hatcher - thurston @xcite and wajnryb @xcite : for every @xmath206 , the complex @xmath196 is connected and simply - connected .
[ thm : ht ] first , hatcher - thurston @xcite and wajnryb @xcite proved that the cut - system complex of a finite - type surface is connected and simply - connected . to see that this is also the case for @xmath196
observe that , for every finite set of vertices @xmath207 , the union of the curves defining the elements of @xmath32 together fill a finite - type subsurface of @xmath2 .
we are finally in a position to prove theorem [ thm : fg ] : as mentioned earlier , we are going to apply theorem [ thm : brown ] to the action of @xmath4 on @xmath196 .
first of all , theorem [ thm : ht ] tells us that @xmath196 is connected and simply - connected .
now , the classification theorem for infinite - type surfaces ( theorem [ thm : class ] above ) and the fact that vertices are defined by a finite set of curves , together imply that @xmath4 acts transitively on the set of vertices of @xmath196 .
since edges and 2-cells of @xmath196 are defined in terms of intersection numbers , and by a finite set of curves , we deduce that @xmath4 acts cocompactly on @xmath196 .
thus , it remains to justify why the stabiliser of a vertex ( resp .
edge ) is finitely presented ( resp .
finitely generated ) .
consider first the case of the stabiliser of a vertex @xmath208 of @xmath209 .
fix an orientation on each curve of @xmath208 . up
to passing to a subgroup of finite index , and abusing notation , we may assume that every element of @xmath210 fixes every curve of @xmath208 together with its orientation .
now , cutting @xmath2 open along the elements of @xmath208 we obtain a surface homeomorphic to @xmath100 , and a short exact sequence @xmath211 where @xmath212 is the group generated by the dehn twists along the elements of @xmath208 . by proposition [ prop : extfk2 ] , we deduce that ( a finite - index subgroup of ) @xmath210 is finitely presented , as desired . using a totally analogous reasoning , we deduce that ( a finite - index subgroup of ) the stabiliser of an edge @xmath213 fits in a short exact sequence @xmath214 therefore , the stabiliser of an edge is finitely presented , and in particular finitely generated .
this finishes the proof of theorem [ thm : fg ] * remark .
the case @xmath149 . * in @xcite
, funar - kapoudjian proved that the group @xmath16 is finitely generated .
however , the methods used here to prove that @xmath4 is finitely presented ( for @xmath3 finite ) do not apply to the case of @xmath16 .
nevertheless , it is plausible that , along the lines of the arguments of @xcite described above , one can find a suitable definition of a _ reduced pants complex _ of @xmath17 for which the action of @xmath16 satisfies the hypotheses of brown s theorem [ thm : brown ] . in the light of this discussion , we close this section with the following question : is the group @xmath16 finitely presented ? [ quest : finpres ]
in this section , we explain the methods of @xcite , and show how to adapt them to prove theorem [ thm : homology ] . recall this result
asserts that the rational @xmath5-th homology of @xmath4 is equal to the stable rational @xmath5-th homology group of the mapping class group of genus @xmath3 , provided @xmath5 is suitably small compared to @xmath3 .
we commence by defining all these terms . as always , denote by @xmath20 the connected orientable surface of genus @xmath3 with @xmath21 boundary components , and
let @xmath23 be its pure mapping class group ; recall that the elements of @xmath23 fix each boundary component pointwise . for @xmath215 ,
gluing a pair of pants to a boundary component of @xmath20 gives rise to an injective homomorphism @xmath216 in particular , one has an injective homomorphism @xmath217 between the corresponding homology groups . in @xcite
, harer proved that the @xmath5-th homology group of @xmath23 does not depend on @xmath21 , provided @xmath5 is sufficiently small with respect to @xmath3 .
the current best bound for what sufficiently small " means is due to boldsen @xcite who , building up on unpublished work of harer @xcite , proved the following : : let @xmath218 , with @xmath13 and @xmath219 . then , the homomorphism is an isomorphism for every @xmath220 .
[ thm : harer ] by lemma [ lem : limit ] , the pure mapping class group @xmath90 is the direct limit of the ( partially - ordered ) set of pure mapping class groups of the family of @xmath94-suited subsurfaces of @xmath2 , where @xmath94 denotes the pants decomposition of @xmath2 underlying the rigid structure . in particular , by theorem [ thm : harer ] ,
@xmath221 for all @xmath86 , provided @xmath26 .
for this reason , we will refer to the homology group @xmath222 as the _ stable _ rational @xmath5-th homology group of the pure mapping class group of genus @xmath3 .
we are finally in a position to give a proof of theorem [ thm : homology ] , whose statement we now recall : theorem [ thm : homology ] let @xmath25 .
for every @xmath26 , the @xmath5-th homology group @xmath27 is isomorphic to the @xmath5-th stable homology group of the pure mapping class group of genus @xmath3 .
we translate the proof of ( * ? ? ?
* theorem 3.1 ) to our setting .
first , the resolution of mumford s conjecture by madsen - weiss @xcite implies that @xmath223 has finite dimension over @xmath224 .
now , we apply the lyndon - hochschild - serre spectral sequence method to the short exact sequence , which recall reads @xmath225 in other words , there is a spectral sequence approximating the homology of @xmath4 whose second page is @xmath226 we now claim that @xmath62 acts trivially on @xmath227 for all @xmath228 . accepting this claim for the moment
, it follows that @xmath229 on the other hand , @xmath62 is rationally acyclic @xcite , and hence the only non - trivial possibility for the right - hand side is @xmath230 . in particular , the spectral sequence for the homology of @xmath4 collapses at the second page , which implies that @xmath231 for all @xmath228 , as desired .
thus it remains to show : * claim .
* @xmath62 acts trivially on @xmath227 .
since @xmath223 has finite dimension over @xmath224 , we get a finite - dimensional representation @xmath232 , for some @xmath233 .
since @xmath62 is simple , it follows that this representation is either trivial or injective . on the other hand ,
every finitely - generated subgroup of @xmath234 is residually finite , while @xmath62 is not . in particular ,
the given representation is trivial .
this finishes the proof of theorem [ thm : homology ] .
in this section , we prove theorem [ thm : ainj ] , whose statement we now recall : theorem [ thm : ainj ] let @xmath36 . if @xmath37 , then there are no @xmath32-injective maps from @xmath38 to @xmath4 .
we remind the reader that , given groups @xmath235 , we say that a homomorphism @xmath236 is _
if for every subgroup @xmath237 with @xmath238 , we have that @xmath239 is injective . first , we observe that @xmath4 has subgroups isomorphic to @xmath240 for all @xmath3 : for every @xmath52 , @xmath4 contains a subgroup isomorphic to @xmath240 .
[ lem : a - inj ] before we give the proof of the lemma , we need some definitions .
first , given a surface @xmath63 ( possibly of infinite type ) , the _ curve complex _
@xmath241 is the flag simplicial complex whose @xmath5-simplices are sets of @xmath242 pairwise disjoint curves on @xmath63 .
now , given a simplicial graph @xmath243 , the _ right - angled artin group _
@xmath244 defined by @xmath243 is the group generated by the vertices of @xmath243 , subject to the relations that two generators commute if and only if they are connected by an edge in @xmath243 . with these definitions
, we have : let @xmath243 be the finite simplicial graph with three vertices and exactly one edge .
it is easy to see that @xmath243 embeds into @xmath245 for all @xmath5 , provided @xmath21 is sufficiently big ; in fact , as long as @xmath246 , this is true for all @xmath21 .
now , ( * ? ? ?
* proposition 7.16 ) implies that the right - angled artin group @xmath244 embeds as a subgroup of @xmath247 .
the result now follows since @xmath244 is isomorphic to @xmath248 , and @xmath4 contains @xmath23 for every @xmath86 , by lemma [ lem : limit ]
. we can now prove theorem [ thm : ainj ] : let @xmath36 , with @xmath7 .
seeking a contradiction , suppose there were an @xmath32-injective map @xmath249 recall the short exact sequence @xmath250 a result of bleak - salazar @xcite and corwin @xcite asserts that thompson s group @xmath62 does not have subgroups of the form @xmath248 .
in particular , since @xmath251 is @xmath32-injective , it gives rise to a non - trivial map ( for @xmath21 sufficiently big if @xmath252 ) @xmath253 since @xmath254 is finitely presented and @xmath90 is a direct limit of finitely presented groups , by lemma [ lem : limit ] , we deduce that there is @xmath255 and a nontrivial map @xmath256 but every such map is trivial , by ( * ? ? ?
* theorem 1.1 ) or @xcite .
hence we get the desired contradiction . as mentioned in the introduction , since every isomorphism is @xmath32-injective , we get the following classification of asymptotic mapping class groups up to isomorphism : corollary [ cor : iso ] let @xmath36
. then @xmath40 if and only if @xmath41 .
the goal of this section is to prove theorem [ thm : autos ] . throughout
, @xmath42 will denote the automorphism group of the group @xmath43 .
the first ingredient in the proof of theorem [ thm : autos ] is the following : if @xmath30 then @xmath257 .
[ prop : autos_pmod ] we note that , prior to this work , this result had been obtained by patel - vlamis @xcite for @xmath258 . before explaining a proof of proposition [ prop : autos_pmod ] , we note the following easy observation : if @xmath30 then @xmath90 is a normal subgroup of @xmath9 .
[ lem : pmod_normal ] let @xmath259 and let @xmath127 be a support for @xmath3 .
consider an arbitrary @xmath260 , noting that any @xmath94-suited subsurface @xmath261 containing @xmath262 is a support for @xmath263 .
moreover , @xmath263 induces the trivial permutation on the set of boundary components of @xmath261 , and therefore @xmath264 . in order to prove proposition [ prop : autos_pmod ] , we follow ivanov s strategy for proving that automorphisms of ( sufficiently complicated ) finite - type mapping class groups are conjugations .
first , we observe : let @xmath30 .
then , every automorphism of @xmath90 sends dehn twists to dehn twists .
[ lem : twists2twists ] we stress that , if @xmath265 , then lemma [ lem : twists2twists ] quickly follows from ( * ? ? ?
* corollary 1.5 ) plus the fact that @xmath90 is a direct limit of @xmath94-suited subsurfaces of @xmath2 ( lemma [ lem : limit ] ) .
we now present a simpler argument , valid for arbitrary @xmath3 , which was suggested to us by juan souto : let @xmath266 .
we take an arbitrary @xmath94-suited subsurface @xmath267 of genus @xmath3 ; in addition , we will assume that @xmath127 has complexity at least 3 ( note that the latter assumption is only relevant if @xmath268 ) .
since @xmath269 is finitely presented and @xmath90 is a direct limit of pure mapping class groups of @xmath94-suited subsurfaces of genus @xmath3 , by lemma [ lem : limit ] , we deduce that there exists a @xmath94-suited subsurface @xmath270 such that @xmath271 .
let @xmath272 be the ( compact ) subsurface of @xmath261 supporting @xmath273 ; in particular , this implies that @xmath274 , and that @xmath275 and @xmath273 have the same centralizer in @xmath90 .
now , observe that , given a compact subsurface @xmath276 , the rank of the center of the centralizer of @xmath277 in @xmath90 is equal to the number of boundary components of @xmath96 . in particular , since @xmath269 and @xmath273 are isomorphic , the discussion of the previous paragraph implies that @xmath272 and @xmath127 have the same number of boundary components .
again , @xcite or @xcite imply that they have equal genus , as @xmath127 has genus @xmath3 and @xmath274 .
in other words , we have deduced that @xmath127 and @xmath272 are homeomorphic .
we claim : _ claim . _
the isomorphism @xmath278 is induced by a homeomorphism @xmath279 . if @xmath127 has genus @xmath280 , then the claim is ( * ? ? ? * theorem 9 ) .
we now sketch an argument valid for arbitrary genus .
the center @xmath281 of @xmath269 is the free abelian group generated by the dehn twists along the boundary components of @xmath127 .
observe that @xmath251 induces an isomorphism @xmath282 and that @xmath283 is isomorphic to the pure mapping class group of the surface which results from gluing a punctured disk to every boundary component of @xmath127 . by the work of numerous authors ( see , for instance , ( * ? ? ?
* theorem 1.3 ) for a comprehensive statement ) , the isomorphism is induced by a homeomorphism @xmath126 between the punctured surfaces . to see that @xmath126 extends to a homeomorphism @xmath279 inducing @xmath251 , we use the well - known _ lantern relation _
( see , for instance , ( * ? ? ?
* section 5 ) ) .
indeed , as the complexity of @xmath127 is large enough , given a boundary component @xmath284 we may find a collection @xmath285 of six curves , all of them essential in @xmath127 , such that the dehn twist along @xmath286 is expressed as a product of ( suitable powers of ) dehn twists along the @xmath287 s , via the lantern relation .
thus the claim follows .
the claim above implies that @xmath278 is induced by a homeomorphism @xmath288 , and in particular sends dehn twists to dehn twists .
the lemma now follows since @xmath127 was arbitrary and every curve on @xmath2 is contained in some @xmath94-suited subsurface of genus @xmath3 .
continuing with the arguments towards a proof of proposition [ prop : autos_pmod ] , we next claim that every automorphism of @xmath90 induces an automorphism of the curve complex @xmath289 .
indeed , let @xmath290 be an automorphism . by lemma [ lem : twists2twists ] , given a curve @xmath291
there exists a curve @xmath292 such that @xmath293 , where @xmath294 denotes the ( left ) dehn twist about the curve @xmath286 .
since @xmath295 if and only if @xmath296 , we get that the curve @xmath292 above is in fact unique , and hence @xmath251 induces a well - defined map @xmath297 by the rule @xmath298 .
observe that the map @xmath299 is simplicial , since two dehn twists commute if and only if the defining curves are disjoint .
moreover , it is bijective , with inverse the simplicial self - map of @xmath300 induced by @xmath301 .
now , hernndez - morales - valdez @xcite proved the following analogue of ivanov s celebrated result @xcite on automorphisms of the curve complex : let @xmath30 .
then the group of simplicial automorphisms of @xmath300 is isomorphic to @xmath302 [ thm : hmv ] we are finally in a position to prove proposition [ prop : autos_pmod ] : let @xmath30 .
we wish to show that the natural homomorphism @xmath303 is an isomorphism .
we first prove that @xmath304 is surjective . to this end , let @xmath305 , and consider the automorphism @xmath306 induced by @xmath251 . by theorem [ thm : hmv ] , there exists a unique @xmath307 with @xmath308 for all @xmath309 .
in fact , up to precomposing @xmath251 with @xmath310 , we may assume that @xmath311 for all @xmath309 .
using again a direct translation of a classical argument in the finite - type setting , we will prove that the automorphism @xmath251 is in fact conjugation by @xmath126 .
to this end , let @xmath312 , and choose an arbitrary curve @xmath151 on @xmath2 .
we have : @xmath313 on the other hand , : @xmath314 combining both equations , we obtain that @xmath315 for every curve @xmath151 on @xmath2 . by the infinite - type alexander method ( see @xcite ) , we deduce that @xmath316 for every @xmath317 . in other words , @xmath251 is the identity on @xmath90 , as desired .
we now embark in the proof of theorem [ thm : autos ] .
the key ingredient is the following unpublished result of kapoudjian : [ lemma - kapou ] let @xmath30 and @xmath318 .
then @xmath319 if and only if for any finite set @xmath320 there exists @xmath321 such that @xmath322 and @xmath323=1 $ ] .
let @xmath171 .
suppose first that @xmath324 , and let @xmath320 be an arbitrary finite set .
choose a minimal support @xmath127 for @xmath126 , so that @xmath325 and @xmath126 is identity outside @xmath127 .
observe that @xmath326 is a support for @xmath327 .
then we may take @xmath328 such that @xmath329 for all @xmath330 , and set @xmath331 , which commutes with @xmath327 . for the opposite direction let @xmath127 be a defining surface for @xmath126 ;
recall this means that @xmath126 sends @xmath127 to @xmath118 , and is rigid in the complement of @xmath332 .
up to enlarging @xmath127 if necessary , we may assume that @xmath127 has at least five boundary components .
we now choose some finite set @xmath320 as in the statement , and which furthermore satisfies that the @xmath333 all send @xmath127 to itself , and that the group generated by the @xmath333 induces every possible permutation of the set of boundary components of @xmath127 . seeking a contradiction ,
suppose that @xmath334 .
then there exists a connected component of @xmath124 whose image by @xmath126 is disjoint from itself .
this implies that every defining surface of any @xmath335 as in the statement must be contained in @xmath127 .
furthermore , the permutation of the set of boundary components of @xmath127 induced by @xmath336 must commute with every element of the conjugacy class of the ( nontrivial ) permutation induced by @xmath126 .
now , if a permutation in a symmetric group on @xmath337 elements commutes with every conjugate of a non - trivial permutation , then @xmath148 is the trivial permutation . in particular
, the permutation induced by @xmath336 is the trivial one , and @xmath338 .
hence @xmath126 belongs to the normalizer of @xmath339 in @xmath4 , contradicting our assumption .
as an immediate consequence we get : for every @xmath30 , @xmath90 is characteristic in @xmath4 .
[ cor : charac ] in the light of corollary [ cor : charac ] , every automorphism of @xmath4 induces an automorphism of @xmath90 , which we have already determined in proposition [ prop : autos_pmod ] .
now , in order to calculate the automorphism group of @xmath4 from this , we will make use of the following small technical result this is surely well - known , but we include a proof for completeness .
let @xmath340 be a normal subgroup of a group @xmath43 , and suppose it has trivial centralizer in @xmath43 .
suppose @xmath341 is an injective homomorphism of @xmath43 such that @xmath342
. then @xmath343 .
[ lem : restriction ] since @xmath344 is the identity on @xmath340 and @xmath340 is normal in @xmath43 we have that , for every @xmath345 and every @xmath346 , @xmath347 since @xmath336 is arbitrary , it follows that @xmath348 belongs to the centralizer of @xmath340 in @xmath43 , and therefore @xmath349 , by hypothesis .
since @xmath126 is also arbitrary , the result follows .
the motivation for lemma [ lem : restriction ] is the following claim : the centralizer of @xmath90 in @xmath9 is trivial .
[ lem : centralizer ] suppose @xmath350 commutes with every element of @xmath90 , and in particular with every dehn twist .
in particular , we have that @xmath351 for every curve @xmath151 on @xmath2 .
but , as we mentioned in the proof of proposition [ prop : autos_pmod ] , the alexander method of @xcite implies that every element of @xmath9 that fixes every curve on @xmath2 is the identity .
we are finally in a position to prove theorem [ thm : autos ] : we first prove that the natural homomorphism @xmath352 is surjective . to this end , let @xmath353 be an automorphism . by corollary [ cor : charac ] , @xmath251 induces an automorphism @xmath354 which is the restriction of @xmath251 to @xmath90 .
proposition [ prop : autos_pmod ] implies that there exists @xmath355 such that @xmath356 is conjugation by @xmath357 , denoted @xmath358 . from the proof of lemma [ lem : twists2twists ] for any ( compact ) @xmath94-suited subsurface @xmath359 of genus @xmath3 , there exists a compact subsurface @xmath360 such that the restriction @xmath361 sends @xmath362 isomorphically onto @xmath277 .
observe that for any @xmath94-suited subsurface @xmath359 the group @xmath363 admits a canonical embedding into @xmath4 , by extending rigidly classes of homeomorphisms of @xmath189 to classes of homeomorphisms of @xmath2 .
this procedure could be extended , but not canonically , to any compact subsurface @xmath276 of genus @xmath3 .
indeed there exists a compact @xmath94-suited subsurface @xmath267 containing @xmath96 . by further adjoining pair of pants to @xmath96 , where needed , one can assume that all components of @xmath364 are pairwise homeomorphic .
we first embed @xmath365 into @xmath366 , and then embed @xmath366 into @xmath4 .
the image of @xmath365 into @xmath4 depends on the choice of @xmath127 and of an element of @xmath367 .
however , we have : for any @xmath94-suited subsurface @xmath359 of genus @xmath3 , the restriction @xmath368 sends @xmath369 isomorphically onto some copy of @xmath370 .
moreover , @xmath368 is conjugacy by @xmath357 .
as @xmath371 is the conjugacy by @xmath357 , we have @xmath372 , when @xmath373 is a boundary circle .
the stabilizer of the set @xmath374 in @xmath4 is the stabilizer of the simplex of @xmath375 determined by the set of curves in @xmath376 , with respect to the @xmath4-action .
we call it the stabilizer of the multicurve .
let @xmath96 be a compact subsurface of @xmath2 and denote by @xmath377 the permutation group on the set of boundary components of @xmath96 .
there are canonical homomorphisms sending a mapping class into the corresponding induced permutation : @xmath378 we denote by @xmath379 the fibred product of the two homomorphisms above , making the following diagram commutative : @xmath380 assume now that @xmath96 has genus @xmath3 .
then the stabilizer of the multicurve @xmath381 in @xmath9 is @xmath379 .
further , the stabilizer in @xmath4 of the multicurve @xmath376 should be sent by @xmath251 into the stabilizer of the multicurve @xmath381 , where @xmath382 . since @xmath363 stabilizes @xmath376 we derive that @xmath383 .
if @xmath384 and @xmath385 denote , respectively , the projections onto the first and second factors of the fibred product then we derive a map
@xmath386 , such that @xmath387 is the conjugation @xmath358 by @xmath357 .
we claim that @xmath388 . by composing with @xmath389
this reduces to show that if a homomorphism @xmath390 restricts to identity on @xmath362 then @xmath391 is the identity .
if @xmath392 sends a boundary component @xmath286 into @xmath195 , then @xmath393 so that @xmath394 . thus the homomorphism @xmath395 induced by @xmath391 is identity .
then the five lemma implies that @xmath391 is injective and lemma [ lem : restriction ] permits to conclude , as the centralizer of @xmath362 into @xmath363 is trivial , if the complexity of the surface @xmath189 is large enough .
further , the image @xmath396 is trivial , since @xmath397 . therefore @xmath398 factors through the quotient @xmath399 , namely @xmath400 , and the map @xmath401 is the isomorphism induced by the conjugation @xmath358 .
this implies that there exists a section @xmath402 which determines an embedding of @xmath365 into @xmath9 so that @xmath403 is the image of @xmath365 .
this lemma shows that the natural map @xmath404 coincides with @xmath358 .
now , an immediate consequence of theorem [ thm : fg ] is that @xmath4 has a system of generators consisting uniquely of elements belonging to @xmath363 for some @xmath94-suited subsurfaces @xmath189 .
in particular , @xmath251 and @xmath358 agree on @xmath4 .
since @xmath358 preserves @xmath4 , @xmath405 , as claimed . to see that the homomorphism @xmath406 is injective , suppose that @xmath407 induces the identity automorphism of @xmath4 . then @xmath357 also induces the identity automorphism of @xmath90 which implies , in particular , that it fixes every dehn twist . at this point ,
corollary [ cor : charac ] tells us that @xmath90 is normal in @xmath4 , since it is characteristic . in addition
, @xmath90 has trivial centralizer in @xmath9 , and therefore in @xmath4 , by lemma [ lem : centralizer ] .
hence , lemma [ lem : restriction ] implies that @xmath357 is identity .
we saw in theorem [ thm : autos ] that the automorphism group of @xmath4 coincides with its normalizer in @xmath9 , as long as @xmath408 .
we are going to explore some basic properties of this normalizer ; as we will see , these represent a certain departure from the case of finite - type mapping class groups .
first , we have : let @xmath409 .
let @xmath43 be a group with @xmath410 .
setting @xmath411 , there is a natural surjective homomorphism : @xmath412 where @xmath413 .
[ lem : rubin ] as @xmath90 is a normal subgroup of @xmath43 , the map @xmath414 that sends every element of the normalizer to the automorphism given by conjugation by that element is a well - defined homomorphism .
further , by a deep theorem of rubin ( see @xcite ) there is an isomorphism @xmath415 finally , any element of @xmath416 lifts to an element of @xmath43 , so that the homomorphism in the statement is surjective .
a recent result @xcite shows that the outer automorphism group @xmath417 of @xmath62 is infinite . combining this with the lemma above
, we obtain : let @xmath175 .
the homomorphism @xmath418 is surjective . in particular
, @xmath48 is an infinite group .
[ cor : out2 ] we remark that a theorem of ivanov @xcite mentioned above asserts that @xmath419 is always a finite group ( in fact , trivial in all but finitely many cases ) .
thus corollary [ cor : out ] represents a limitation in the dictionary between asymptotic and finite - type mapping class groups . finally , we prove that the homomorphism of corollary [ cor : out ] has infinite kernel : let @xmath30
. then the kernel of the homomorphism @xmath418 is infinite .
consider first the surface @xmath106 , which recall that is homeomorphic to a sphere minus the union of a cantor set and an isolated puncture .
using a totally analogous argument to that of lemma [ lem : rubin ] , we deduce that there is a surjective homomorphism @xmath420 let @xmath94 be the pants decomposition underlying the rigid structure on @xmath106 , and consider the element @xmath421 obtained as the product of all ( say left ) half - twists about the elements of @xmath94 .
we claim that @xmath422 . to see this , observe that if @xmath127 is a defining surface of an element @xmath423 , then each connected component of @xmath424 is preserved by @xmath425 ; moreover , this element acts as the identity on every such component . in particular , @xmath426 and
thus @xmath422 .
furthermore , the short exact sequence implies that @xmath427 unless @xmath428 , since @xmath429 fixes every end of @xmath106 but does not have compact support unless @xmath428 .
hence it provides an example of an infinite order element in @xmath430 . on the other hand ,
its image under the homomorphism is obviously trivial , and thus this homomorphism has infinite kernel .
after all this discussion , we treat the case of @xmath2 .
to do so , we simply embed @xmath106 in @xmath2 in such way that the pants decomposition @xmath94 underlying the rigid structure of @xmath106 is sent to the pants decomposition underlying the rigid structure of @xmath2 . using the same arguments as above , we deduce that the image of @xmath431 under the homomorphism induced by the embedding @xmath432 produces an infinite - order element in the kernel of the homomorphism @xmath418 .
we start this section by proving theorem [ thm : t ] , which is a direct consequence of the existence of the short exact sequence . before we prove the result , recall that a discrete group @xmath43 is said to have _
kazhdan s property ( t ) _ if every action of @xmath43 by continuous affine isometries on a real hilbert space has a fixed point . on the other hand
, @xmath43 has the _ haagerup property _ if it admits a proper action by continuous affine isometries on a hilbert space .
it follows that if @xmath43 has both properties , then @xmath43 is finite .
we refer the reader to the book @xcite for a thorough exposition of these properties
. we may now prove theorem [ thm : t ] , whose statement we now recall : theorem [ thm : t ] let @xmath52 .
then @xmath4 does not have property ( t ) .
let @xmath52 , and recall from equation that there is a surjective homomorphism @xmath170 .
it is known ( see , for instance , ( * ? ? ?
* theorem 1.7.1 ) ) that property ( t ) is preserved by passing to a quotient . on the other hand
, a result of farley @xcite asserts that @xmath62 has the haagerup property .
since @xmath62 is infinite , it follows that @xmath62 does not have property ( t ) , and thus neither does @xmath4 .
we now proceed to give a proof of theorem [ thm : lattice ] .
the proof is again based on properties of the group @xmath62 , plus a result of farb - masur @xcite which asserts that every homomorphic image of a lattice in a finite - type mapping class group is finite . more concretely , they proved : let @xmath49 be an irreducible lattice in a semisimple lie group of real rank at least two .
for every @xmath433 , every homomorphism @xmath434 has finite image .
[ thm : fm ] before giving a proof of theorem [ thm : lattice ] , we remind the reader of its statement : theorem [ thm : lattice ] let @xmath49 be a lattice in a semisimple lie group @xmath43 of real rank at least two , where @xmath43 has no compact factors isogenous to @xmath50 or @xmath51 . for every @xmath52 ,
any homomorphism @xmath53 has finite image .
let @xmath43 be a lie group as in the statement .
as such , @xmath43 has property ( t ) ( see ( * ? ? ?
* theorem 13.2.4 ) , for instance ) , and therefore so does @xmath49 ( see ( * ? ? ? * theorem 1.7.1 ) ) .
for such a lattice @xmath49 , let @xmath435 be a homomorphism .
recall again the short exact sequence : @xmath436 as in the proof of theorem [ thm : t ] , the fact that @xmath62 has the haagerup property implies that @xmath437 is finite .
but @xmath438 is also finite . to see this ,
observe that since @xmath49 is finitely presented ( see ( * ? ? ?
* theorem 4.7.10 ) ) , lemma [ lem : limit ] implies that there exists a finite - type subsurface @xmath63 of @xmath2 such that @xmath438 is contained in @xmath88 . at this point ,
theorem [ thm : fm ] tells us that @xmath438 is also finite , and the result follows .
in this section we will prove theorem [ thm : nonlinear ] , which asserts that @xmath4 does not admit an injective homomorphism into @xmath55 for any @xmath21 , provided @xmath3 is finite . as mentioned in the introduction
, we will do so by showing that @xmath4 contains a copy of _ thompson s group @xmath57_. recall that @xmath57 is the group of piecewise - linear self - homeomorphisms of @xmath439 $ ] that preserve rational dyadic numbers , are differentiable except at finitely many dyadic rationals , and at each interval of differentiability the slopes are ( positive ) powers of 2 .
a strongly related group which we will also need is _
thompson s group @xmath144 _ , whose definition is the same as that of @xmath57 but with the unit circle @xmath440 instead of @xmath439 $ ] .
one has the well - known inclusions @xmath441 ; for a proof , as well as a detailed discussion on thompson s groups , see @xcite .
we now prove : [ split ] let @xmath30 . the short exact sequence , that is @xmath442 splits over thompson s group @xmath57 .
the result is known for @xmath145 ( see @xcite ) , where in addition it is shown that the sequence @xmath15 splits over thompson group s @xmath144 , identified with the subgroup of @xmath15 preserving the whole visible side of @xmath92 .
consider now a disk punctured along a cantor set , and denote it @xmath443 .
the surface @xmath443 comes equipped with an obvious rigid structure that comes from that of @xmath92 under the natural subsurface embedding @xmath444 .
we regard @xmath443 as properly embedded into an annulus @xmath445 $ ] , so that the cantor set is the dyadic cantor subset of @xmath446 .
we fix an identification of @xmath144 with the group of orientation - preserving , piecewise affine - dyadic ( i.e. with dyadic break points and dyadic slopes ) homeomorphisms of @xmath447 preserving the cantor set .
denote by @xmath448 be the group of asymptotically rigid mapping classes of @xmath443 which preserve the visible side .
then @xmath448 is a central extension of @xmath144 by the cyclic group @xmath449 generated by the dehn twist along the boundary curve of @xmath443 .
let @xmath391 be a homeomorphism of @xmath443 representing the class @xmath450 .
observe that the universal cover @xmath451 is properly embedded into the strip @xmath452 $ ] .
let @xmath453 be the lift of @xmath391 to @xmath451 that is identity on the boundary line @xmath454 , and note that the restriction of @xmath455 only depends on the class @xmath456 .
this provides an isomorphism between @xmath448 and the group of increasing piecewise - affine dyadic homeomorphisms of @xmath457 which commute with the translation @xmath458 .
the latter group is the central extension of @xmath144 associated to the euler class ( see ( * ? ? ?
* theorem d ) ) , which is known to split over @xmath57 . finally , we prove theorem [ thm : nonlinear ] : let @xmath30 . by proposition
[ split ] , @xmath459 contains an isomorphic copy of thompson s group @xmath57 .
now , @xmath57 has simple commutator subgroup @xcite , and in particular is not residually finite .
since finitely - generated subgroups of linear groups are residually finite and @xmath57 is finitely generated , the result follows .
in this section we introduce a close relative of @xmath4 , namely the group @xmath8 of _ half - twists _ , introduced in @xcite in the case @xmath145 ( under the name @xmath60 therein ) . by definition , @xmath8 properly contains @xmath4 , and below we will prove that @xmath8 is still finitely presented .
nevertheless , we will see in theorems [ thm : limit ] and [ thm : limit2 ] that the main source of interest for studying the group @xmath8 is that it serves as a finitely - presented approximation of @xmath9 .
indeed , we will show that @xmath8 is dense in @xmath9 . before continuing we
remark that , while we will prove that @xmath8 satisfies a number of properties analogous to the ones we established for @xmath4 in the previous sections ( for instance , theorems [ thm : half - finpres ] and [ thm : half - autos ] below ) we will not be able to show that they all hold true .
the main obstacle in this direction is that , although @xmath8 will fit in a short exact sequence ( see below ) analogous to the short exact sequence for @xmath4 , the group that will play the rle of thompson s group @xmath62 is less studied than @xmath62 is .
it seems plausible , however , that once this hurdle has been removed , all of our main results for @xmath4 will hold for @xmath8 as well .
the _ signed thompson s group _ is the group of right - continous bijections of the unit circle that map the set of dyadic rationals to itself , are differentiable except at finitely many points , and such that , on every interval of differentiability , they are affine maps whose derivatives are ( positive or negative ) powers of 2 . in the light of the definition above ,
the group @xmath460 is closely related to thompson s group @xmath62 . by analogy with the equivalent definition of @xmath62 in terms of trees ( compare with section [ sec : v ] ) , every element of @xmath460 is encoded by an _ enhanced symbol _
@xmath461 , where @xmath462 are admissible trees , @xmath463 is a bijection and @xmath464 , up to leaf equivalence as defined in section [ sec : v ] .
let @xmath409 , and fix a rigid structure on @xmath2 .
the group @xmath8 is the group of isotopy classes of _ asymptotically quasi - rigid _ ( as in definition [ def : asrig1 ] ) homeomorphisms @xmath465 . as proved in @xcite , @xmath59 is the group of @xmath466-diffeomorphisms of the sphere minus a certain cantor set .
write @xmath469 for the group of isotopy classes of asymptotically quasi - rigid homeomorphisms of @xmath100 , which recall is the surface obtained from @xmath92 by removing @xmath195 isolated points . by using a minor adaptation of the argument in proposition [ prop : extfk2 ]
, we get : as was the case in the proof of proposition [ prop : extfk2 ] , the reduced pants complex of @xmath100 is a connected and simply - connected 2-complex on which @xmath471 acts cocompactly .
moreover , the stabilizer in @xmath471 of every cell is a finite extension of the corresponding stabilizer in @xmath139 ; in particular , it is finitely presented . at this point ,
brown s theorem [ thm : brown ] yields that @xmath471 is finitely presented .
next , all dehn twists are conjugate in @xmath15 .
then the normal subgroup in @xmath471 generated by one dehn twist coincides with the group generated by all dehn twists , namely with @xmath90 .
hence , adding to a finite presentation of @xmath471 one more relator corresponding to a dehn twist yields a presentation of @xmath460 , which is therefore finitely presented . to see that @xmath460 is simplie , first recall that a sufficiently transitive permutation group generated by elements of small support has simple commutator ( see @xcite ) .
the group @xmath460 inherits the sufficient transitivity from the @xmath62-action on the cantor set ( see e.g. @xcite , section 3 ) . on the other hand @xmath460
is generated by the images of half - twists and @xmath62 .
this implies that @xmath460 is generated by elements of arbitrarily small support .
observe that all half twists are conjugate in @xmath59 and hence in @xmath460 , so that @xmath472 has order at most 2 .
now , the class of ( the image of ) a half - twist in @xmath460 is the restriction of an affine orientation - reversing self - homeomorphism of an interval .
it follows that a half twist of an interval is the composition of the permutation exchanging its two halves and the product of the half twists of each halve .
this relation implies that @xmath460 is perfect and hence simple .
the group @xmath8 acts on the cut - system complex @xmath196 , which is connected and simply - connected by theorem [ thm : ht ] .
since @xmath4 acts cocompactly on @xmath196 and @xmath467 , it follows that the action of @xmath8 on @xmath196 is also cocompact .
next , using the same reasoning as in the proof of theorem [ thm : fg ] , the @xmath8-stabiliser of a vertex ( resp .
edge ) of @xmath196 is , up to finite index , the group @xmath473 ( resp @xmath474 ) , which is finitely presented by proposition [ prop : half - finpres ] . at this point
, the result follows again from brown s theorem [ thm : brown ] .
indeed , the the obvious translation of kapoudjian s lemma [ lemma - kapou ] to this setting tells us that @xmath90 is a characteristic subgroup of @xmath8 . at this point
, the proof proceeds word by word as in section [ sec : proofaut ] .
as it turns out , theorem [ thm : limit ] will be a consequence of a slightly stronger result , namely theorem [ thm : limit2 ] below . before we state it , we need some preliminaries .
let @xmath30 and fix , once and for all , a rigid structure on @xmath2 .
recall from section [ sec : rigid ] that the rigid structure is given by a separating curve on @xmath2 , which will be denoted @xmath477 for reasons that will become apparent below , plus a rigid structure on the planar component @xmath478 of @xmath479 .
let @xmath94 be the pants decomposition underlying the rigid structure on @xmath478 .
there is an obvious rooted tree @xmath144 associated to @xmath94 , whose vertices are the curves of @xmath94 , with the root being @xmath477 . as such , every vertex of @xmath144 is naturally labelled by a word @xmath480 in the free semigroup @xmath481 generated by the two letters @xmath482 ( left ) and @xmath483 ( right ) .
we denote by @xmath484 the curve of @xmath94 labelled by the word @xmath485 , and write @xmath486 for the half - twist about @xmath484 . in turn
, this labelling induces a labelling of the set of pairs of pants of @xmath478 by words in @xmath481 .
indeed , we set @xmath487 to be the unique @xmath94-suited pair of pants of @xmath478 that has @xmath484 as boundary component and is contained in the planar component of @xmath488 .
finally , define @xmath489 to be the unique compact subsurface of @xmath2 whose interior is homeomorphic to @xmath490 , and whose boundary curve is @xmath477 .
in addition , for each @xmath491 , let @xmath492 be a @xmath94-suited ( therefore compact ) subsurface whose interior is contained in @xmath493 , and which has @xmath484 as a boundary component .
observe that the choice is far from unique ; however , for any such choice , the set @xmath494 is _ proper _ , in the sense that every compact subsurface of @xmath2 intersects only finitely many elements of this set . because of this
, we will refer to any set @xmath495 as above as a _ proper sequence of subsurfaces_. 1 .
@xmath498 is an element of @xmath499 , 2 .
@xmath500 denotes the length of the word @xmath480 , 3 .
@xmath501 for all @xmath480 , 4 .
the product is ordered lexicographically , and is defined from left to right .
consider @xmath477 . by the classification of infinite - type surfaces up to homeomorphism @xcite ( theorem [ thm : class ] above )
, there exist a ( compact ) @xmath94-suited subsurface @xmath489 , and an element @xmath504 such that @xmath505 sends @xmath477 to itself .
we say that @xmath505 _ straightens _ @xmath477 .
next , consider the curves @xmath506 and @xmath507 .
again , there exists a @xmath94-suited planar subsurface @xmath508 and an element @xmath509 such that @xmath510 sends the set @xmath511 to itself .
therefore , up to precomposing with the half - twist @xmath512 about @xmath477 if necessary , we may in fact assume that @xmath506 ( resp .
@xmath507 ) is sent to itself .
we say that @xmath510 straightens the pair of pants @xmath513 .
we continue this process inductively to find , for all @xmath21 , a proper sequence of ( compact ) @xmath94-suited subsurfaces @xmath492 and elements @xmath514 such that , setting @xmath515 the mapping class @xmath516 straightens every pair of pants @xmath487 for @xmath517 . by lemma [ lem : proper ] , the sequence @xmath518 converges to an element @xmath519 , in such way that @xmath520 fixes @xmath477 and every element of @xmath94 .
moreover , by composing with an appropriate power of a dehn twist at each step , we may assume that @xmath520 induces the trivial twist on each component of @xmath94 .
therefore , @xmath520 is an element of @xmath521 . in other words , we have shown that , up to precomposing @xmath522 with the inverse of this element , the sequence @xmath503 converges to @xmath126 in @xmath9 , as desired .
collin bleak , peter cameron , yonah maissel , andrs navas , feyishayo olukoya , the further chameleon groups of richard thompson and graham higman : automorphisms via dynamics for the higman groups @xmath524 , arxiv:1605.09302 .
c. kapoudjian and v. sergiescu , an extension of the burau representation to a mapping class group associated to thompson s group @xmath144 . in geometry and dynamics , contemp
soc . , providence , r.i .
, 2005 , 141164 . | for @xmath0 , we study a subgroup @xmath1 of _ asymptotically rigid _
mapping classes , for @xmath2 a closed orientable surface of genus @xmath3 minus a cantor set .
the group @xmath4 contains the mapping class group of every compact orientable surface of genus @xmath3 with non - empty boundary .
we show that @xmath4 is finitely presented and that its @xmath5-th homology group coincides with the stable homology of the mapping class group of genus @xmath3 , for @xmath5 small enough with respct to @xmath3 .
we then proceed to prove that the family of groups so obtained enjoys a number of properties analogous to known results for finite - type mapping class groups . more concretely , there are no ( weakly ) injective maps @xmath6 if @xmath7 , every automorphism of @xmath4 is geometric , and every homomorphism from a higher - rank lattice to @xmath4 has finite image .
in addition to the translation of classical results about finite - type mapping class groups , we will show that @xmath4 is not linear , and does not have kazhdan s property ( t ) . finally , in the appendix we will study the group @xmath8 of _ half - twists _ of @xmath2 , which sits between @xmath4 and @xmath9 . more concretely
, we will prove that @xmath8 is finitely presented , and that it is dense in @xmath9 . |
in the last decade many papers were devoted to calculations of the low energy properties of generalized parton distributions ( gpds ) , @xcite .
the main tool to obtain the low - energy expansion of matrix elements is effective field theories based on spontaneous symmetry breaking , such as chiral perturbation theory ( chpt ) or heavy baryon chiral perturbation theory ( hbchpt ) .
investigation of the chiral structure of a hadron is an important task . only in this way
one can obtain the quark mass dependence and the momentum transfer dependence for gpds in a model - independent way .
investigation of the chiral expansion gives us the behavior of a parton distribution at large impact parameters , where it is governed by pion cloud , see .
e.g. @xcite . while the pion parton distributions were examined comprehensively , the nucleon gpds were considered only as mellin moments .
this article is devoted to the exploration of the chiral corrections to the nucleon gpds in the @xmath2-space ( @xmath2 is the longitudinal hadron momentum fraction carried by a parton ) .
there is a principal difference between the chiral expansions of a parton distribution @xmath3 in @xmath2-space and its mellin moments @xmath4 .
the point is that the mellin transformation does not preserve the chiral order of expansion .
let us dwell on it .
the expressions for @xmath3 and @xmath4 are connected unambiguously via forward and inverse mellin transformations @xmath5 using an effective field theory one gets the first terms of the gradient expansion of mellin moments : @xmath6 , where @xmath7 ( @xmath8gev is the character scale of the chiral expansion ) . the remainder of expansion is of order @xmath9 : @xmath10 .
let us consider the function @xmath11 , which is a result of the inverse mellin transform for @xmath12 .
the general statement is that the deviation @xmath13 is not necessarily @xmath14 .
the order of the function @xmath15 depends on @xmath2 and can not be guessed without investigation of the analytical property of the full function @xmath3 . in the case of qcd
perturbative expansion , the deviation @xmath15 is known to be of the same order as a reminder everywhere except very small @xmath2 . for the gradient expansion the situation is more sharp , because the variable @xmath2 is of kinematic origin and expansion over kinematic parameter can significantly change the analytic properties of a parton distribution .
it is clear that the order of deviation @xmath15 increases with the decreasing of @xmath2 .
it follows from the fact that singularities of @xmath3 are concentrated in the region @xmath16 and for large @xmath2 they are shadowed by the singularity at @xmath17 .
indeed the analysis of the pion parton distribution in the chpt @xcite shows that the deviation @xmath15 is @xmath14 for @xmath18 , but became @xmath19 at @xmath20 . in order to reach the declared accuracy @xmath14 at @xmath20 , the contribution of particular class of diagrams should be added .
the calculation of the low-@xmath2 contribution to the pion gpd was performed with the help of recursive equation technic @xcite , and leads to evident effects @xcite . for the nucleon target such
an analysis was never done .
the starting point of the analysis is the operator definition of gpd .
gpds are defined as matrix elements of the light - cone qcd operator .
the light - cone separation @xmath21 is dimensional , and hence has its own counting rule in the chiral expansion .
contribution of some particular region of @xmath21-integration is distinct for different @xmath2 .
this allows one to correlate @xmath21-counting rules with an appropriate @xmath2-region of gpd . in the mellin moments one integrates over all range of @xmath2 and
therefore mixes up distinct @xmath21-counting rules . in order to resolve
this ambiguity , it is convenient to transfer the counting prescription of @xmath21 onto some other object , which is not washed out during mellin transformation .
the natural object is the vector of light cone direction @xmath22 ( @xmath23 ) : first of all , it accompanies the parameter @xmath21 in the gpd operator , second , the normalization of the vector @xmath24 can be changed without changing of gpd definition : @xmath25 therefore , adjusting the special counting rules for the vector @xmath24(@xmath26 ) one obtains the expression for @xmath12 such that @xmath27 in the corresponded area of @xmath2 .
the detailed analysis of the counting rules for vector @xmath24 in different regions of @xmath2 is presented in the sec.ii .
notice , that applying the special counting rules for vector @xmath24 the contributions of the operators with higher number of derivatives to mellin moments can be allowed , contrary the standard counting rules .
then the expression for @xmath12 is not truncated by @xmath28 term , but can contain terms with higher orders of @xmath28 .
this exceedance of the accuracy is needed for the restoration of the proper analytic properties of the distribution .
it also shows that the calculation directly with nonlocal operators looks preferably . in the later case
one should not care on the special counting rules because the parameter @xmath21 is not integrated out , and its size can be directly taken into account .
the price is the incorporation of the nonlocal operator in the gradient expansion .
additional difficulty arises from to the power counting problem of the pion - nucleon chpt .
the standard tool to avoid the counting problem is to use hbchpt , which is a non - relativistic limit of chpt .
however , naive attempt to consider a light - cone operator in the non - relativistic framework is condemned to fail .
the infrared logarithms of straightforward hbchpt are replaced by some cumbersome functions , which often have artificial singularities in @xmath2 .
the reason is the non - commutativity of the non - relativistic limit and the infinite momentum frame nature of the light - cone operator .
therefore , one should consider the nonlocal operator in the covariant framework with consequent procedure of the redefinition of the higher terms by large - mass corrections @xcite .
we notice that the heavy baryon reduction via the infrared regulation @xcite , does not suite here as well , because the loop - integrals with nonlocal operator have infrared properties distinct from the properties of usual loop - integral .
we describe the non - contradictory heavy baryon reduction for loop - integral with nonlocal operators in sec .
iii . in this paper
we calculate the leading non - analytical chiral correction to nucleon gpds at @xmath0 directly from nonlocal operators in the relativistic covariant framework using extended on - mass - shell subtruction scheme , eoms @xcite .
the structure of the article is following : in sec.ii we derive the effective nonlocal operators and show the difference between our operator and the operators constructed in hbchpt . in sec .
iii we perform the detailed investigation of the heavy baryon limit for the nonlocal operators , and derive rules for self - consistent nonlocal operator consideration within chpt .
the results of our calculations are presented in sec.iv .
calculations with lagrangian ( [ complete_l ] ) lead to the power - counting problem , since the nucleon mass is of the same order as the normalization scale of chpt , @xmath41 gev .
the most popular solution of the counting problem is provided by hbchpt @xcite .
the nucleon field is split onto soft and hard components : @xmath42 where @xmath43 is the nucleon velocity vector , @xmath44 .
one eliminates the component @xmath45 using the equations of motion and obtains hbchpt lagrangian .
the pion - nucleon part of hbchpt lagrangian reads @xmath46 where @xmath47 is the spin operator , @xmath48 . in the limit @xmath53 , the parameter @xmath54 . in this article we put @xmath55 to zero , by setting @xmath56 .
the effects related to non - zero @xmath55 can be of the leading order in @xmath57 expansion , and they will be considered in a separate article . in the limit @xmath58
the gpds are @xmath59 where @xmath60 and @xmath61 are the parton distribution functions ( pdfs ) with transverse momentum dependance .
gpd @xmath62 can not be obtain in the limit @xmath56 . in this subsection
we construct the effective vector and axial light - cone operators for the nucleon .
our procedure is very close to one presented in @xcite with minor redefinitions .
the procedure of the matching qcd light - cone operators to the operators in the chiral effective theories is well - known .
one should introduce the nonlocal operator with light - cone separation @xmath21 and transformation properties of the initial operator .
we start with qcd operators : @xmath63 these operators transform in mixed representation of the chiral rotations group .
it is convenient to deal with the left- and right - handed light - cone operators : @xmath64 which transform as the define representation .
the fields involved in the chiral lagrangian transforms under local @xmath65 as : @xmath66 where @xmath67 and @xmath68 are matrices of right and left chiral rotations , and @xmath69 is a matrix of chiral transformations for nucleon field containing field @xmath31 .
the lowest order operators contain no derivatives .
the most general structures respecting the symmetries are : @xmath70 where @xmath71 and @xmath72 .
the integration area over @xmath73 and @xmath74 is @xmath75 , these limits are provided by additional requirement of polynomiality . the functions @xmath76 represent generation functions for the tower of low energy constants , and have the meaning of the double distributions for the nucleon gpds with given isospin number in the chiral limit @xcite .
the properties of operators ( [ o],[o5 ] ) under parity transformations demand that @xmath77 and @xmath78 .
thus , the expression for vector and axial light - cone operators are @xmath79 where the scalar and pseudo - scalar combinations of pion fields are @xmath80 in expressions ( [ o_in_chpt],[a_in_chpt ] ) one has freedom to choose the normalization of the generating function in a convenient way . to recover that the @xmath81 are the double distributions in the chiral order one should calculate the matrix element @xmath82 at the tree level , and compare the result with definition ( [ v_qcd ] ) .
we obtain : @xmath83 where the symbol @xmath84 denotes the chiral limit of a quantity . the similar consideration for function @xmath85
gives : @xmath86 in the limit @xmath0 , which is of the special interest in this paper , the @xmath73-dependence of the operators can be moved out by the translation .
the resulting generating functions are @xmath87 where @xmath88 and @xmath89 are the nucleon pdfs in the chiral limit .
the normalization of the pdfs follows from the vector and axial current operators , it reads : @xmath90 the reduction of the operators ( [ o_in_chpt],[a_in_chpt ] ) to the heavy - baryon form is straightforward : @xmath91 @xmath92 we remind that in hbchpt nucleons are taken to be non - relativistic , and nucleon momenta which flow through the diagrams are reduced momenta : @xmath93 .
the expressions for operators ( [ o_in_hbchpt],[a_in_hbchpt ] ) have the exponential factor @xmath94 , which restores the reduced nucleon momentum to the complete nucleon momentum in the operator vertex .
in principal , this factor can be absorbed to the definition of the low - energy generating functions @xmath81 , but this would lead to the inconvenient scale of parameters @xmath74 and @xmath73 .
we keep the exponential factors as they are , in order to have the simple interpretation ( [ f_interpretation0]-[f - normalization ] ) for the generating functions .
the matrix element of operators ( [ o_in_chpt]-[a_in_chpt ] ) contributes to the gpds @xmath95 and @xmath96 at the tree order , and to all set of gpds at one - loop order .
the leading contribution to the gpds @xmath97 and @xmath62 is given by the operators of the higher order in chiral counting .
the operator giving a chiral limit of gpd @xmath97 is proportional to @xmath98 , and it does not give a contribution to @xmath95 and @xmath96 at tree and one - loop level .
therefore , in this article , which is mostly devoted to gpds @xmath95 and @xmath96 we leave such kind of operators out .
pure pion nonlocal operators respecting the transformation properties of ( [ o],[o5 ] ) can be constructed .
their form and properties were derived at @xcite .
the pure pion operators for the vector and axial pdfs have the form : @xmath99},\end{aligned}\ ] ] @xmath100},\end{aligned}\ ] ] where the integration area for @xmath73 and @xmath74 is @xmath75 , and @xmath101 .
the axial pion operator gives no contribution at @xmath56 ( up to one - loop level ) , and therefore , we will skip it in current consideration .
the low - energy generation function @xmath102 belongs to the pion gpd in the chiral limit via the expressions : @xmath103},\end{aligned}\ ] ] where @xmath104 stands for the isospin , and @xmath105 . for the forward limit of the pion pdf we will use the notation @xmath106 .
the normalization of pion pdfs is : @xmath107 the mellin moments of gpds are widely used in phenomenology .
the operator of the mellin moment can be found using the well - known relation : @xmath108 calculation of the mellin moments in effective field theory is straightforward and was performed in @xcite up to second chiral order .
the result of calculation has the form : @xmath109 where @xmath28 is the chiral expansion parameter , @xmath110 .
the reminder term of the expansion is @xmath111 .
however , the function @xmath112 ( here , @xmath3 is the full parton distribution and @xmath11 is the reconstructed from @xmath12 parton distribution ) is not necessarily @xmath14 in the full domain of @xmath2 .
the main reason for non - commutation of the perturbative expansion and the mellin transformation is the violation of the analytical properties of @xmath3 by the chiral expansion .
there are no doubts , that the low - energy expansion changes the analytical structure of the parton distributions , and the inverse mellin transformation of the low - energy expansion has different significance in different regions of @xmath2 .
the left - hand - side of equation ( [ op<>m_qcd ] ) contains the dimensional parameter @xmath21 , which has its own chiral counting .
the different areas of @xmath21 are responsible for the physical content of the different regions of @xmath2 . performing mellin transformation
one integrates over all range of @xmath2 and therefore mixes up various counting rules for @xmath21 . in order to resolve this ambiguity , it is convenient to refer the counting prescription of @xmath21 to the light - cone vector @xmath22
. this can be done via changing the normalization factor @xmath113 of the vector @xmath24 in eqn.([norm_n ] ) .
adjusting special counting rule to the vector @xmath24 we obtain the expression for @xmath12 such that @xmath114 in the corresponding area of @xmath2 .
general form of @xmath11 reads : @xmath115 where the function @xmath116 can not be expanded in @xmath28 without distraction of the analytical structure . as it will be demonstrated later , often the mellin convolution in ( [ q_form ] ) is well - behaved in total , but it is divergent for every term of taylor expansion of @xmath116 at @xmath117 . performing mellin transformation of ( [ q_form ] ) one gets : @xmath118 where @xmath119 can be further expanded over @xmath28 .
the term @xmath120 is equal to @xmath121 and , therefore , the expression ( [ demonstration of m ] ) coincides with the expression ( [ m_form ] ) at order @xmath14 .
the exceedance of the accuracy in ( [ demonstration of m ] ) is needed for the restoration of the proper analytic properties of the distribution @xmath11 .
we start the consideration of the counting rules for @xmath21 from the analysis of the standard counting rules of hbchpt , where the light - cone vector @xmath22 has no special prescription .
the standard counting rules of hbchpt reads : @xmath122 where @xmath60 is the parameter of gradient expansion , i.e. small momenta or the pion mass , @xmath123 . in the hbchpt
nonlocal nucleon operators also contain the exponent with the heavy nucleon mass , which suppresses soft derivatives .
therefore , in the leading order the hbchpt operator ( [ o_in_hbchpt ] ) reads : @xmath124 taking into account that @xmath125 and ( [ f - interpretation ] ) we found @xmath126 operator ( [ o_in_(25 ) ] ) allows one to obtain only such corrections to pdf those do not touch the intrinsic dynamics of the nucleon .
in other words using rules ( [ standard_counting_rules ] ) one indirectly uses the approximation @xmath127 , where @xmath128 is the form - factor with corresponding quantum numbers .
the standard counting rule @xmath129 relates to the region of @xmath2 close to unity .
that is , at @xmath130 the reconstructed parton distribution @xmath11 deviates as @xmath14 from the full one . in this region the short @xmath21 ( @xmath131 )
dominates in the integral ( [ op<>m_qcd ] ) . in the region
@xmath132 we need to set the condition @xmath133 ( this follows from the condition that the argument of the exponent should be of order unity , otherwise integral is suppressed ) . as a result
one has the following counting rules : @xmath134 and similar for the nucleon field .
also the consistent counting rules should be attached to the vector @xmath43 @xmath135 in this case all light - cone derivatives are of the same importance .
the operators for mellin moments can be obtained directly by the expansion of the operators ( [ o_in_hbchpt],[a_in_hbchpt ] ) .
numerically this area corresponds to @xmath136 . in this region
the loops with composite operators should be calculated with a great care , because usually ultraviolet logarithms of @xmath137 can become infrared . and , indeed , such effects take a place and we discuss them in the next section .
the region @xmath138 is governed by the higher derivative terms since @xmath139 and @xmath140 .
the integral ( [ op<>m_qcd ] ) is controlled by the huge light - cone distance @xmath141 .
therefore , the higher terms of the chiral expansion are of the same importance as the leading order and should be taken into account simultaneously .
the terms belonging to this region have the form of @xmath142-functions and their derivatives , see the case of pion @xcite .
the resummation of such series in all orders of chiral expansion leads to the leading term , which describes the long - range behavior of the pion cloud .
this mechanism is also known as the `` chiral inflation '' , see @xcite .
we stress , that many problems with intermediate regions can be avoided performing calculation directly with nonlocal operators . keeping the parameter @xmath21 alive
one leaves the decision of the interesting @xmath2-region until the final result . as a price one has to perform more involved calculation .
there are five diagrams contributing to the leading order of pdf chiral expansion , fig.1 .
the calculation of these diagrams is almost straightforward .
however , the final result contains some uncertainties , related to non - commutativity of the heavy baryon limit and light - cone definition of the operator , which we discuss in this section . to show features of the calculation with nonlocal operators in heavy baryon theories we consider in details the diagram b. for concreteness
, we consider the vector isovector operator .
the diagram b is interesting because it mixes the axial and the vector pdfs with each other .
such a mixing is believed to be strongly suppressed , because the mellin moments of this diagram are zero at leading order , due to convolution of the nucleon spin - vector with the nucleon velocity . whereas the calculation with nonlocal operator gives non - zero result already at the leading order .
the straightforward calculation of loop integral with the lorentz invariant action ( [ complete_l ] ) before the renormalization procedure reads : @xmath143}^{-\epsilon } \delta(x-\eta\beta){\right]},\end{aligned}\ ] ] where @xmath144 , and @xmath145 is a parameter of dimensional regularization : @xmath146 .
the integral in the brackets gives : @xmath147}^{-\epsilon } \delta(x-\eta\beta ) = 2(m^2)^{1-\epsilon}~\frac{\bar y}{\beta } { \left[}\bar y^2+\alpha^2y{\right]}^{-\epsilon}\theta(0<y<1),\end{aligned}\ ] ] where @xmath148 and @xmath149 .
this part is proportional to the large mass and can not be considered as a correction without additional adaptation .
the standard approach of the heavy baryon theory posses the rescaling of the nucleon fields ( [ field_rescaling ] ) and afterwards expansion in the small ratio @xmath150 . in the language of the feynman diagrams it means a partial resummation of the perturbative expansion and
it was never done explicitly . the bypass way was suggested in @xcite : the heavy baryon result is provided by the infrared singular part of the loop - integral , whereas the infrared regular part is to be subtracted into the higher order coupling constants .
this method works perfectly for usual loop - integrals , but faces problems with loop - integrals of type ( [ b1 ] ) . applying formally the method of @xcite to the integral ( [ b1 ] ) one gets :
@xmath151 } & & b_{\text{\cite{becher:1999he}}}(x)= \frac{g_a}{4}\frac{\bar n\tau^a\hat n n}{p_+}\gamma(-1+\epsilon ) \\{\nonumber } & & \times \int_{-1}^1 d\beta{\stackrel{o}{\delta q}}(\beta ) { \left[}\delta(x-\beta ) \frac{(m^2)^{1-\epsilon}}{(4\pi f_\pi)^2}+\frac{2(m^2)^{1-\epsilon}}{(4\pi f_\pi)^2}\frac{\bar y}{|\beta|}(\bar y^2+\alpha^2 y)^{-\epsilon}\theta(y<1){\right]}.\end{aligned}\ ] ] this expression is still proportional to @xmath152 .
the reason is that the @xmath142-function of the nonlocal operator changes the low - energy behavior of the loop - integral and does not allow to reach the infrared pole at @xmath153 .
also expression ( [ b[17 ] ] ) contains incorrect @xmath154-function , which results to the singular integral over @xmath74 .
the integration over @xmath2 for expression ( [ b[17 ] ] ) is infrared divergent , due to the incorrect @xmath154-function .
this reproduces the infrared singular part for the loop - integrals for the mellin moments .
the proper way to reduce the integral ( [ b11 ] ) to the heavy baryon limit is not to extract the infrared singular part of the loop - integral , but to subtract its infrared regular part .
the regular part of the loop integral is not changed by the nonlocality , because at the limit @xmath58 the delta - function can be moved out of the loop - integral . therefore , the reduction to the heavy baryon limit for nonlocal operators consists in the subtraction of the pole at @xmath58 , with the subsequent renormalization procedure .
the described procedure is known as extended on - mass - shell renormalization @xcite . after the renormalization procedure for the expression ( [ b1 ] )
reads : @xmath155}.\end{aligned}\ ] ] this expression is @xmath156 , in spite of the large mass @xmath32 in the second term .
the expression ( [ b5 ] ) has a general form ( [ q_form ] ) .
expanding the logarithm in the second term in powers of @xmath73 one get the asymptotic series , because every term starting from @xmath157 is divergent at @xmath158 integral due to the factor @xmath159 .
this is a bright example of the non - commutativity of the large mass limit with the light - cone formulation of the nucleon structure operators .
the mellin moment of the expression ( [ b5 ] ) reads : @xmath160 where @xmath161 @xmath162 the leading term of the expansion for @xmath163 in powers of @xmath73 vanishes due to cancelation of integrals over logarithm and delta - function .
therefore , the non - analytical part mellin moment of the mixing parity term is @xmath164 this result coincides with the results of @xcite .
the analytical part of the diagram is of order @xmath156 and accompanied by the leading contribution of higher chiral order operator .
the mixing of vector and axial pdfs , which is of order @xmath165 as follows from the mellin moment calculation , is of order @xmath166 . in the isoscalar case , where this diagram is enforced by the chiral algebra factor ,
the correction is of order @xmath167 for @xmath168 .
we have calculated the leading order non - analytical corrections to the nucleon parton distributions @xmath169 , @xmath170 and @xmath171 .
the results were obtained in the lorentz invariant theory and then reduced to the heavy baryon limit , as it is described in the previous section . here
we present the full set of results depending on the isospin and on the lorentz structure .
we use the notation @xmath172 for the reminder of the expansion .
it indicates that the calculated by us terms corresponds to the leading logarithmical contribution for the mellin moments , although they contains the non - logarithmical functions .
the finite part of expression contains the tree expression for the higher derivative gpd operators .
they are not presented here . for the interested reader we present diagram - by - diagram listing in the appendix a. in order to present the result in more convenient way , we use the following notations : for the vector pdfs @xmath173 @xmath174 and for the axial pdf @xmath175 where @xmath176 and @xmath177 are the nucleon pdfs in the chiral limit , and @xmath178 is the pion pdf in the chiral limit .
the subscript @xmath104 stands for the isospin .
the coefficient functions @xmath179 , @xmath180 and @xmath181 are listed below . for the vector pdf the functions @xmath179
are : @xmath182 for the axial pdf the functions @xmath180 are @xmath183 the parity - mixing functions @xmath184 are the same for the axial and vector case : @xmath185 the pure pion contribution to the vector pdfs is given by : @xmath186 the coefficient functions @xmath181 defining the vector pdf @xmath187 are following @xmath188 where @xmath189 @xmath190 the presented results can be compared with existent in literature results for mellin moments .
the moments @xmath191 of parton distributions @xmath192 and @xmath193 diagram - by - diagram reproduce the vector and axial form factors given in @xcite .
also , our results reproduce the mellin moments of @xmath192 and @xmath193 presented in @xcite and corresponded nucleon gravitational form factors given in @xcite .
the expressions for the distribution @xmath187 have some disagreement with the results of @xcite .
the disagreement arises from the nonlocal operators with one derivative , which should be taken into account for the distribution @xmath187 , but are not considered in this article .
the last two terms in ( [ qend-1]-[qend ] ) have a special behavior in @xmath2 . at @xmath130
these terms contribute to the parton distribution @xmath192 as @xmath157 , but at @xmath194 their contribution increases and becomes of order @xmath73 . in the region @xmath195 these terms give contributions of order @xmath196 .
the fact that corrections change their chiral order with @xmath2 is a character feature of the pure pion sector .
similar terms in nucleon - pion sector ( [ c0]-[dc1 ] ) do not change the order of correction with the changing of @xmath2 .
also we are drawing attention to the @xmath142-function contributions in expressions ( [ qend-1],[qend ] ) .
these contributions have the special nature , because at @xmath195 they are formally of the same order as the tree term . together with similar contributions from the higher orders of chiral expansion they form the leading chiral correction to the nucleon pdf in the region @xmath197 .
the operator evidence of the these terms was presented in sec.2.2 .
the resummation of the @xmath142-function contributions in all orders of chpt is an important task .
it leads to the model - independent picture of the nucleon at large impact parameters .
using nucleon - pion chpt we have calculated the leading non - analytic chiral corrections for nucleon pdfs with non - zero momentum transfer in @xmath2-space .
we have restricted ourselves to the case @xmath56 and to the gpd operators of the leading order .
we have derived the general rules to deal with light - cone nonlocal operators for the theories with heavy particles .
our method is the generalization of eoms scheme ( the extended - on - mass - shell subtraction scheme by @xcite ) and it can be applied for operators with various quantum numbers .
application of another methods for solving the hierarchy problem , such as hbchpt and infrared regularization scheme @xcite , result to artificial singularities in the physical domain of @xmath2 .
the mellin moments of our calculations coincides with the known results .
however , our analysis shows that the mellin moments calculated within standard hbchpt approach can be used for reconstruction of the original parton distribution only in the region @xmath130 , and fails to restore the distribution already at @xmath198 .
our results can be used for investigation of the distributions down to @xmath199 .
we found that the derived corrections give the visible contribution of order @xmath200 .
we have shown that there exist a parity - mixing term , which mixes vector and axial pdfs .
the mixing contribution is of order @xmath9 , which can not be deduced from consideration of the mellin moments . because of uncertainties in the inverse mellin transformation the mixing terms were believed to be of order @xmath201 , which two orders of magnitude less .
the pion contributions to the vector gpd contain the term proportional to @xmath202 , which is a bright sign of necessity to sum up such terms in all orders of the chiral expansion .
they must be resummed at @xmath195 , since all @xmath142-terms are of the same order as the leading term .
the resummation can be done in the leading logarithmical order with the help of the recursive equations @xcite .
the result gives the leading asymptotic term in the long distance behavior of nucleon gpds , see e.g.@xcite .
the slope of pdfs is related to the average impact parameter squared as @xmath203 we have found , that the contribution of the pion sector of pdfs given by ( [ qend-1]-[qend],[sp ] ) is larger in magnitude then the contribution of the nucleon sector ( [ c0]-[tildac1],[sn ] ) for @xmath204 , which is in a good agreement with general chiral picture of the nucleon @xcite .
for example , average squared impact parameter @xmath205 for the vector isovector parton distribution @xmath206 is : @xmath207 where the first term comes from nucleon contribution @xmath208 , whereas the second term comes from the pion contribution @xmath209 .
the important in phenomenology case of non - zero @xmath210 , as well as complete calculation of gpds @xmath97 and @xmath62 , also can be considered in the presented framework and will be covered in further publication .
we would like to thank m.polyakov , n.kivel and g.gegelia for numerous stimulating discussions and real interest to our work .
also we thank p.wein for comments .
this work is supported by bmbf grant 06bo9012 .
in this appendix we present the listing of the diagram contributions , in @xmath2-representation and their mellin moments at the leading logarithmical order .
the diagrams are shown in fig.1 . to obtain the results ( [ q0]-[qend ] ) the renormalization of the nucleon field should be added .
diagrams @xmath211 corresponding to the matrix element ( [ v_qcd ] ) can be presented as the following convolution : @xmath212}u(p),\end{aligned}\ ] ] where @xmath213 and @xmath214 belong to the nucleon parton distributions @xmath169 and @xmath171 , respectively .
the vector pdf @xmath215 should be replaced by the axial pdf @xmath216 for the diagram @xmath217 , and by the pion pdf @xmath218 for the diagrams @xmath219 and @xmath97 .
three diagrams @xmath225 contributing to the axial matrix element ( [ a_qcd ] ) will be presented in a similar convolution : @xmath226 where @xmath227 belongs to the nucleon parton distribution @xmath170 .
the axial pdf @xmath216 should be replaced by the vector pdf @xmath215 for the diagram @xmath217 .
@xmath228 below we present non - analytic contributions to mellin moments of @xmath192 , @xmath193 and @xmath187 nucleon parton distributions at order @xmath156 in the chiral expansion .
we define mellin transformation of the functions @xmath229 , @xmath230 and @xmath231 as @xmath232 and give a list of their moments diagram - by - diagram : @xmath233 @xmath234 @xmath235 functions @xmath236 and @xmath237 for zeroth mellin moment diagram - by - diagram reproduce results of @xcite for the vector and axial form factors .
functions @xmath238 differ in the logarithm parts .
n. kivel , m. v. polyakov and s. stratmann , nucl - th/0407052 .
m. diehl , a. manashov and a. schafer , eur .
phys . j. a * 31 * ( 2007 ) 335 [ hep - ph/0611101 ] .
m. diehl , a. manashov and a. schafer , eur .
j. a * 29 * ( 2006 ) 315 [ hep - ph/0608113 ] .
s. -i .
ando , j. -w .
chen and c. -w .
kao , phys .
d * 74 * ( 2006 ) 094013 [ hep - ph/0602200 ] .
m. dorati , t. a. gail and t. r. hemmert , nucl .
phys . a * 798 * ( 2008 ) 96 [ nucl - th/0703073 ] . m. diehl , a. manashov and a. schafer , phys .
b * 622 * ( 2005 ) 69 [ hep - ph/0505269 ] . n. kivel and m. v. polyakov , hep - ph/0203264 .
n. kivel and m. v. polyakov , phys .
b * 664 * ( 2008 ) 64 [ arxiv:0707.2208 [ hep - ph ] ] .
n. kivel , m. v. polyakov and a. vladimirov , phys .
d * 79 * ( 2009 ) 014028 [ arxiv:0809.2064 [ hep - ph ] ] .
m. strikman and c. weiss , phys .
d * 80 * ( 2009 ) 114029 [ arxiv:0906.3267 [ hep - ph ] ] .
m. alberg and g. a. miller , phys .
* 108 * ( 2012 ) 172001 [ arxiv:1201.4184 [ nucl - th ] ] .
n. kivel , m. v. polyakov and a. vladimirov , phys . rev .
* 101 * ( 2008 ) 262001 [ arxiv:0809.3236 [ hep - ph ] ]
. i. a. perevalova , m. v. polyakov , a. n. vall and a. a. vladimirov , arxiv:1105.4990 [ hep - ph ] . m. r. schindler , j. gegelia and s. scherer , phys .
b * 586 * ( 2004 ) 258 [ hep - ph/0309005 ] .
e. e. jenkins and a. v. manohar , phys .
b * 255 * ( 1991 ) 558
. + v. bernard , n. kaiser , j. kambor and u. g. meissner , nucl .
b * 388 * ( 1992 ) 315 .
a. v. radyushkin , phys .
d * 58 * ( 1998 ) 114008 [ hep - ph/9803316 ] . | within the pion - nucleon chiral perturbation theory we derive the leading chiral correction to the nucleon gpd at @xmath0 .
we discuss the difficulties of consideration of nonlocal light - cone operators within the theory with a heavy particle and the methods to solve the difficulties .
the consideration of the chiral corrections directly for nonlocal operators allows to resolve the ambiguity of the inverse mellin transformation .
in particular , we show that the mixing between axial and vector gpds are of order @xmath1 , which is two orders of magnitude less that it follows from the mellin moments calculation . |
the first observational detection of a bump associated with a `` cosmological '' burst came with afterglow observations of grb 980326 ( bloom et al .
there , the early afterglow was seen to fade rapidly in the first few days ( @xmath2 ) with a fairly generic afterglow spectrum ( @xmath3 ) . at day 22 and day 28
, imaging and spectroscopic detections revealed a source that was about 60 times brighter than the early extrapolation and significantly more red ( @xmath4 ) . by the next series of observations at 200 days ,
the bump had faded by at least a factor of ten .
hst imaging observations later revealed a faint galaxy at the position of the transient ( fruchter et al .
2001 ) . what powered the bump in grb 980326 ?
the timescale for rise and decay are natural in the supernova interpretation .
moreover , the red color of the bump can be understood as due to metal line blanketing of a core - collapsed supernova redward of @xmath14000 in the restframe .
that the spectrum differed significantly from the early afterglow spectrum suggests a different physical emission process than in the afterglow .
in other words , the bump in 980326 disfavors a synchrotron shock origin , such as might be expected from a more dense external medium or from delayed energy injection of a central source . bumps , attributed to sne , have since been claimed in a number of other cosmological bursts .
the bump in the afterglow of grb 970228 was based on observations in at least three bandpasses ( reichart 1999 ; galama et al .
2000 ) . moreover , the probable turn over in the spectrum at @xmath5 m ( reichart 2001 ) could not be easily explained by thermal emission from dust or a dust echo .
the peak of redshifted thermal emission should occur redward of @xmath12 @xmath6 m and dust echo spectra should not exhibit a strong roll over .
optical / ir bumps discovered 2030 days after grbs 000911 ( lazzati et al .
2001 ) , 990712 ( bj " ornsson et al . , 2001 ) , 980703 ( holland et al .
2001 ) provided suggestive evidence of bumps all of which the authors claimed found a reasonable explanation in the context of an added supernova component . for several bursts , however , no bump was detected to limits of comparable brightness to 1998bw ( see price et al .
2002 for a review ) .
note that due to severe line blanketing in the restframe uv , a bump from a supernovae is not expected to be seen at optical wavelengths for bursts beyond redshifts of @xmath7 whereas bumps from dust echoes could be seen to higher redshifts .
consistent with the sn hypothesis , no bumps were detected by our group in the afterglows of grbs 010222 ( @xmath8 ) and 000926 ( @xmath9 ) .
the ability to detect the bump in 980326 was aided by two important occurrences .
first , the rapid fading of the early afterglow effectively removed any possible light contamination from the afterglow beyond a few days .
second , the host galaxy was exceedingly faint ( @xmath10 mag ) .
therefore , as we pointed out in bloom et al .
( 1999 ) , a bump in new grbs could be easily outshone for bright host galaxies and/or by bright afterglows .
the latter difficulty can be minimized by observing bursts with rapidly fading afterglow or with early temporal breaks ( presumably due to jetting ) . even for these , without an excellent multi - wavelength characterization of the early - time afterglow , which allows for accurate predictions of the flux at 1040 days , bump detections will be more difficult to claim unambiguously .
overcoming the former host difficulty requires either an intrinsically faint galaxy ( like 980326 ) or the ability to resolve the grb afterglow from its host . to dig deeper into the bump luminosity function , observations with the _ hubble space telescope _
, then , seemed the natural choice .
as of this workshop there have been 12 reported searches for bumps in grb afterglows for bursts originating from @xmath11 . of these ,
half utilized only ground - based observations , and half used both ground - based and hst photometry .
grb 020405 ( price , this workshop ) and grb 020331 ( soderberg , this workshop ) are two new positive detections from our hst program .
as illustrated in figure 1 , for the bursts where a bump was seen or an upper - limit found from ground - based observations alone , the host galaxy is a dominant contaminant .
interestingly , only three bumps appear brighter than their respective hosts at the same wavelengths .
hst observations have allowed us to probe nearly four magnitudes fainter than the total magnitude of the host ( in the case of grb 010921 ; @xmath12 ) .
the non - detection of a bump in grb 010921 is the deepest constraint on a bump : any bump must have been at least 70% fainter than 1998bw at the redshift of the burst . in other words , despite the added depth with hst photometry , _ we still have not been able to probe for bumps fainter than about _
@xmath13 _ mag_.
given the low redshift of grb 011121 and extensive ground - based observations at early times , we undertook a multi - epoch multi - wavelength program with hst to try to detect a bump and test our hypothesis that any detected bump could be due to a supernova .
garnavich et al .
( 2002 ) first pointed out that the @xmath14-band flux at day 14 was higher than the extrapolated light curve from earlier times .
as described in more detail in bloom et al .
( 2002 ) , we detected a significant bump in four hst filters . in that paper
we described how the light curve and spectra over the next 78 days appeared to resemble a supernova .
in addition , afterglow models of 011121 provided the first clear - cut case of a wind - stratified medium around the burst .
this , coupled with the sn interpretation , provides strong evidence for a massive star origin . as both we and garnavich et al .
( 2003 ) described , the behavior of the bump of 011121 did not follow a simple redshifted 1998bw : this rise time appeared a to be quicker ( 17% more rapid ) and the peak flux lower ( 55% fainter ) .
a few explanations could be possible .
the supernova could be of a different type than 1998bw ( e.g. , type iin ; garnavich et al .
, 2003 ) . instead , the supernova could be a type ic but with an energy and ejecta mass between that of 1998bw and 1994i ( see figure 1 ) .
alternatively , the supernova could be a dimmed version of 1998bw but with an explosion date that preceded the grb by @xmath15 days ( @xmath16/dof @xmath17 ; zeh et al .
, this workshop ) .
accepting the supernova interpretation , the zeh analysis is an important one because it places constraints on the relative timing of the supernova and the grb . in a `` supranova ''
, a supramassive neutron star is created as the massive star explodes to produce a supernova ( vietri & stella 1998 ) .
when this neutron star spins down ( via magnetic breaking or gravitational radiation ) losing centrifugal support it collapses to form a black hole producing the grb .
thus a supernova could precede a grb in time .
unfortunately , the time offset is not a strong prediction of the supranova hypothesis in fact , we point out that the original incarnation of the model suggested a 10 year offset , which is clearly ruled out by the 011121 observations .
the short time - offset in 011121 ( consistent with contemporaneous grb and supernova explosions ) underscores one of the more fundamental observational limits in understanding the origin of bumps and grb progenitors . since the explosion time of even the best studied supernovae
can not be dated to better than about @xmath18 days , we may never be able to infer the true time offset for short - delay supranovae .
that is , we may never be able to distinguish between the collapsar model ( zero delay ) and short - delay supranova scenarios based upon optical observations alone ( neutrino arrival times would help ! ) .
clearly not all bumps can be fit with redshifted versions of 1998bw , suggesting an intrinsic diversity in the properties of bumps ( even without appealing to a supernova interpretation ) .
if these bumps are due to supernovae then this diversity is unsurprising : local examples of core - collapsed supernovae are anything but standard .
mazzili et al .
( 2002 ) compiled spectra and light curves of a few well - observed core - collapse supernovae and emphasized the large scatter in peak fluxes ( @xmath19 to @xmath20 mag ) and broad range of colors at peak . since the best non - detection of a bump reaches only @xmath21 mag , i suggest that we may be only detecting the tip of the iceberg of supernovae associated with grbs . 1 . * late - time bumps are real and common .
* bumps have been detected with high significance in at least five grb afterglows ( 980326 , 970228 , 020405 , 011121 , 020331 ) with several other proposed bump identifications ( e.g. , 000911 , 990712 , 980703 ) .
the non - detections or ambiguity in the significance of bumps have been largely due to contaminating light from the host galaxies .
hst observations have allowed us to push bump sensitivity almost four magnitudes fainter than the integrated host magnitude .
still , the best non - detection reaches only to @xmath22 mag . 2 . *
the best case suggests a supernova origin .
* grb 011121 is the best case for a bump and has a spectrum and light curve similar to a type ic or iin supernova .
this is also the best example of a burst that occurred in a wind - stratified medium .
* limits on supranovae . * the original supranova hypothesis , which posited an offset in time between the grb and the associated supernova of 10 years , is clearly ruled out .
modified - supranovae ( with time offsets of less than a @xmath1one week ) are stilled allowed but may be indistinguishable from collapsar models using bump observations alone . with such a small number of intensive bump searches , we know very little about the frequency of occurrence of grb bumps .
in a supernova - centric interpretation , the current non - detections of bumps at redshifts @xmath11 may be due to a lack of sensitivity and intrinsic supernovae diversity . despite the concordance of a number of detected bumps with simplistic _ a priori _ supernovae models , a spectrum of a bump showing either the absence or presence of metal lines would be most convincing one way or another .
( somewhat lost in the various disagreements over the specifics of the bump emission and timescales is : most viable alternatives put forth thus far to explain bumps require a massive star progenitor rather than merger products . ) in the future , to more accurately characterize the nature of bumps , we hope to undertake hst spectroscopy of a bump ( see figure 2 ) and determine if bump decay rates are indeed exponential . , g. , _ et al .
_ 2001 apj ( _ letters _ ) , 552 , l121 , j. s. _ et al .
_ 1998 , apj ( _ letters _ ) , 506 , l105 , j. s. _ et al .
_ 1999 , nature , 401 , 453 , j. s. _ et al .
_ 2002 , apj ( _ letters _ ) , 572 , l45 , z. g. and lu , t. 1998 , a&a , 333 , l87 , a. a. and blandford , r. 2000 , apj ( _ letters _ ) , 534 , l151 fruchter , a. , vreeswijk , p. , and nugent , p. 2001 , gcn notice 1029 , t. j. _ et al .
_ 1998 , nature , 395 , 670 , t. j. _ et al .
_ 2000 , apj , 536 , 185 garnavich , p. m. , _ et al .
_ 2002 , gcn notice 1273 garnavich , p. m. , _ et al .
_ 2003 , accepted to apj , s. _ et al .
_ 2001 , a&a , 371 , 52 , d. _ et al .
_ 2001 , a&a , 378 , 996 mazzali , p. a. _ et al .
_ 2002 , apj ( _ letters _ ) , 572 , l61 , p. a. _ et al .
_ 2002 , accepted to astrophysical journal letters , d. e. 1999 , apj ( _ letters _ ) , 521 , l111 . 2001 , apj , 554 , 643 , j. c. , matzner , c. d. , and mckee , c. f. 2001 , apj , 551 , 946 , m. and stella , l. 1998 , apj ( _ letters _ ) , 507 , 45l , e. and draine , b. t. 2000 , apj , 537 , 796 woosley , s. e. 1993 , apj , 405 , 273 | observations of late - time optical bumps have been reported for several grbs .
the timescale for such bumps , and colors of such when available , find a natural explanation as due to associated supernovae .
ground - based and hst observations of the afterglow and bump of grb 011121 , in particular , place the strongest constraints yet on the physical nature of the supernova and any alternative explanations , such as supranovae or dust echoes .
i summarize the search for underlying bumps in other grbs and make the case for the supernova hypothesis in light of observed bumps and bump non - detections ( e.g. , grb 010921 ) .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in there is a good deal of theoretical motivation to expect to see bumps in grb afterglow lightcurves . for massive star progenitors ( woosley 1993 )
, a bump can arise either by reflection / reprocessing of the afterglow light by surrounding dust ( waxman & draine 2000 ; esin & blandford 2000 ) or from a supernova that accompanies the grb .
a bump might also arise from delayed energy injection by the central source ( dai & lu 1998 ) .
if the afterglow encounters a shell of material with higher density at @xmath0 cm , then the afterglow could rebrighten . for compact binary mergers ( e.g. , double neutron star coalescence ) ,
a late - time bump is not a natural expectation since , a ) mergers should occur in homogeneous regions of low density , b ) explosive nucleosynthesis leading to a supernova is not expected around such systems , and c ) the creation of a stable neutron star after merger that is capable of re - injecting energy after @xmath120 days seems implausible .
d. lazzati ( this workshop ) has given an excellent overview of various progenitor scenarios .
thus the mere existence of bumps offer a strong discriminator between merger and massive star scenarios . within the confines of the massive star progenitors model
, the details of the bumps and accompanying afterglows should also offer insight into the specifics of the progenitors .
for example , a bump from a supernova should have the spectral and temporal characteristics of other supernovae observed in the local universe ; woosley ( 1993 ) suggested that baryon contamination of the relativistic jets could be minimized if this supernova was of type ic .
a bump from thermal dust emission could have similar rise and fall timescales as a supernova , but the spectrum would be thermal and featureless , without any of the metal - line blanketing seen in supernovae .
a bump from dust echoes could mimic supernovae timescales ( reichart 2001 ) , but the bump should fade as a power - law rather than an exponential . -0.55 |
recently , canuto & mazzitelli ( 1991 , cm ) proposed an improved model for stellar convection . being derived from a turbulence model , it takes into account the contribution of the full spectrum of the turbulent convective eddies , to the convective flux . in stellar interiors
the microscopic viscosity is very small compared to the turbulent viscosity , implying that the turbulent spectrum spans many decades in wavenumber space .
therefore , in this respect , the cm model represents a significant improvement over the mixing length theory approach ( mlt ) which is a one eddy ( the largest ) approximation to the spectrum .
moreover , in the cm model , the turbulent mixing length scale is the depth @xmath2 , so there is no need for an adjustable free parameter like the mlt @xmath3-parameter .
the resulting convective fluxes are higher than those of the mlt for high convective efficiencies , and smaller than them for low efficiencies .
the model performs better than the mlt when applied to stellar structure ( dantona & mazzitelli , 1994 ; mazzitelli , dantona & caloi 1995 , stothers & chin 1995 , althaus & benvenuto 1996 ) , helioseismology ( baturin & mironova 1995 , monteiro et al .
1995 , antia & basu 1995 ) and stellar atmospheres ( kupka 1995 ) . in the cm model , the turbulence spectral function @xmath4 is determined by the timescale controlling the energy input from _ buoyancy _ , that is , the _ source _ that generates the turbulence .
this timescale is expected to depend on the parameters of the source as well as on the _ turbulence spectrum _ itself .
the quantification of the latter dependence , within the cm model , is far from obvious .
thus , cm ( 1991 ) assumed that the above timescale can be approximated by the inverse of the growth rate of the unstable modes of the _ linearized _ equations . by construction ,
the latter depends only on the source and is _ independent _ of the _ turbulence _ it generates .
the linear rate was used also by canuto goldman and chasnov ( 1987,cgc ) who , generalizing the work of canuto and goldman ( 1985 ) , proposed a model for fully developed turbulence .
the linear rate was also employed by hartke canuto & dannevik ( 1988 ) in the framework of a dia ( direct interaction approximation ) model for turbulent convection and by canuto , cheng , hartke , & schilling ( 1991 ) for edqnm ( eddy damped quasi - normal markovian ) models .
the rate of energy input in the cm model must be improved since a fully developed turbulence is expected to regulate the energy input from source ( buoyancy ) .
the lack of feedback from the turbulence on the energy input , prevents the model from being selfconsistent .
however , because of the complex structure of the edqnm closure , the implementation of a selfconsistent rate into the formalism of the cm model is not simple .
thus , we were forced to simplify the structure of the nonlinear interactions , so as to be able to formulate a workable selfconsistent treatment . in modeling the nonlinear interactions
, we follow the work of cgc ( 1987 ) .
however , in the present model we generalize the definition of the eddy correlation timescale , thus correcting some shortcomings in the physics involved , and leading to an improved closure .
the resulting effective rate of energy input from the source ( buoyancy ) depends now on both the _ source _ and the _ turbulence_. as stated , the present approach is complementary to that of the cm model . here , the focus is on the selfconsistent rate of energy input and not on the closure , which is much simpler than that of the cm model . in spite of the highly nonlinear structure of the model equations , it is possible to solve them _ directly _ with no need for iterations .
this reduced considerably the amount of numerical work required , and allowed for an exploration of the model for a wide range of values of the convective efficiency , @xmath5 , defined in equation ( 51 ) . for each value of @xmath5
we obtained the spectral function which determines the turbulence bulk quantities .
thus , we computed , as functions of @xmath5 , the convective flux , the turbulent viscosity , the turbulent mean squared velocity and the root mean squared turbulent pressure .
we have applied the new model to the main sequence evolution of a solar model as well as to the evolutions of an extreme pop ii chemical composition ( y=0.23 , z=@xmath6 ) stars with @xmath7 @xmath8 .
the results are generally similar to those of the cm model
. however , the new model has the advantage that the overshoot required to fit the solar model is much smaller , in accord with recent empirical estimates , and the ages of globular clusters are also somewhat smaller than the corresponding ages in the cm model ( which in turn are smaller than those derived within the mlt framework ) .
let us consider a fully developed stationary turbulent convection characterized by the spectral functions @xmath9 and @xmath10 , of the turbulent velocity , temperature fluctuations and the turbulent convective flux , respectively .
before doing so , we write the dynamical time - dependent equations obeyed by these spectral functions ( yamaguchi 1963 )
. @xmath11 @xmath12 and @xmath13 where @xmath14 , @xmath15 , and @xmath16 denote the nonlinear transfer terms for the turbulent velocity , temperature , and convective flux . here
, @xmath17 is the gravitational acceleration , @xmath3 is the coefficient of thermal volume expansion at constant pressure ( equaling @xmath18 for an ideal gas ) , @xmath19 is the superadiabatic temperature gradient , @xmath20 is the microscopic viscosity , and @xmath21 is the microscopic thermometric conductivity appearing in the expression for the conductive flux @xmath22 in stellar interiors , the dominant conductive flux is the radiative flux and thus @xmath21 is the radiative thermometric conductivity .
the function @xmath23 is given by @xmath24 with @xmath25 measuring the anisotropy of the eddy corresponding to the wavenumber @xmath26 . here
, @xmath27 and @xmath28 stand for the horizontal and vertical wavenumbers , respectively ( the vertical direction is that of the gravitational acceleration ) . in equation ( 1 )
, @xmath29 plays the role of the energy source driving the velocity fluctuations .
more precisely , it equals the rate of energy per unit mass and per unit wavenumber , fed to the turbulence velocity field at wavenumber @xmath26 .
the term @xmath30 is the rate of energy per unit mass and unit wavenumber dissipated at @xmath26 by the microscopic viscosity , while @xmath31 represent the rate of energy per unit mass and unit wavenumber transferred to wavenumbers _ other _ than @xmath26 .
analogous interpretations apply to equation ( 2 ) which describes the temperature fluctuations field . as stated above , we are interested in stationary turbulence .
thus , in equations ( 1)(3 ) we set the time derivatives of the spectral functions equal to zero .
following cgc ( 1987 ) , we assume that @xmath32 which equations ( 1 ) , ( 2 ) , and ( 3 ) show to be equivalent to the assumption : @xmath33 for equation ( 7 ) to be satisfied , the velocity and temperature fluctuations , at any @xmath26 , must be in phase .
this is expected to be the case for turbulent convection where the temperature plays the role of an _ active scalar _ whose fluctuations _ drive _ the velocity fluctuations .
we adopt the simplifying assumption that the transfer terms in equations ( 1 ) and ( 2 ) describe transfer from small to large wavenumbers only ( from large spatial scales to small ones ) .
while justified in three dimensional turbulence where energy flows predominantly from large to small scales , this simplifying assumption neglects non - local interactions ( in @xmath26 space ) that give rise to reverse transfer ( backscatter ) .
denoting by @xmath34 the wavenumber corresponding to the largest eddy , we shall represent the transfer as @xmath35 where @xmath36 is the mean squared turbulent vorticity and @xmath37 is the turbulent viscosity at wavenumber @xmath26 , exerted by all eddies of smaller size ( larger @xmath26 ) .
as such , it is expressed as an integral with limits @xmath26 and @xmath38 .
we shall return later to the definition of the integrand .
similarly , @xmath39 where @xmath40 is the analog to the mean squared turbulent vorticity and @xmath41 is the turbulent conductivity at wavenumber @xmath26 exerted by all eddies of smaller size ( larger @xmath26 ) . with these closures ,
we obtain by integrating equations ( 1 ) and ( 2 ) @xmath42 and @xmath43 the last two equations allow us to express the spectral function @xmath44 in terms of @xmath45 .
we obtain @xmath46 with @xmath47 @xmath48 where a prime denotes a differentiation with respect to @xmath26 . substituting equation ( 14 ) in equation ( 7 ) yields @xmath49 thus
, once @xmath45 is known so are @xmath44 and @xmath10 , implying that one needs to solve only for the spectral function @xmath45 . in order to do so ,
let us return to equation ( 12 ) and express @xmath10 in terms of @xmath45 using equation ( 17 ) .
we obtain @xmath50 where @xmath0 is the shorthand abbreviation @xmath51 whose dimension are inverse time .
the combination @xmath52 plays , in equation ( 18 ) , the role of the inverse of the timescale controlling the energy input into the turbulence at wavenumber @xmath26 from the source ( buoyancy in the present case ) .
more specifically , the _ net rate of energy input _ from the source per unit mass and unit wavenumber , at @xmath26 , is @xmath53 .
it is obvious , from equation ( 19 ) , that @xmath0 depends on the _ source _ ( buoyancy ) and on the _ turbulent state_. thus , it conforms with the requirements discussed in 1 . in what follows
we show that equations ( 18 ) and ( 19 ) allow us to express @xmath0 in terms of the turbulent viscosity @xmath37 .
first , differentiate equation ( 18 ) with respect to @xmath26 .
the result is @xmath54 next , use equations ( 15 ) , ( 20 ) and the definition of y(k ) , equation(9 ) , to get @xmath55 with the help of equation ( 19 ) , we transform equation ( 21 ) into a second order algebraic equation for @xmath56 : @xmath57 whose positive solution is @xmath58 we adopt a relation between @xmath41 and @xmath37 @xmath59 so that , as in cgc ( 1987 ) , @xmath60 equals the constant @xmath61 for large pecle numbers , @xmath62 , while @xmath63 , for small @xmath64 . using equation ( 24 ) to express
@xmath65 and @xmath66 in terms of @xmath67 , equation ( 23 ) now takes the form @xmath68 with @xmath69 and @xmath70 with this @xmath56 , @xmath0 of equation ( 19 ) becomes @xmath71 thus , we succeeded in expressing @xmath0 in terms of the turbulent viscosity @xmath37 , even though the latter _ is still unspecified_. in order to solve equation ( 18 ) ( or equivalently equation ( 20 ) ) we need an additional relation between the turbulent viscosity @xmath37 and the spectral function @xmath45 . without loss of generality , @xmath37 can be written as @xmath72 where @xmath73 has the dimensions of an inverse time . in order to determine @xmath73 let us focus on its physical meaning .
differentiation of equation ( 29 ) yields that the turbulent viscosity contributed by eddies in the wavenumber interval @xmath74 is @xmath75 thus , @xmath73 is proportional to the inverse of the eddy - correlation timescale at wavenumber @xmath26 , or heuristically , the inverse of the timescale for the eddy breakup .
the eddy is damped because of two processes : interaction with the turbulent viscosity and the interaction with the source ( microscopic viscosity is not considered here since in stars @xmath76 ) .
one can envisage the eddy as being `` scattered '' by `` collisions '' with the smaller eddies ( turbulent viscosity ) and by the source that drives energy into the eddy .
the effective rate for the breakup ( the inverse of the correlation timescale ) will be taken to be the sum of the rates for the two processes as if they were operating independently : @xmath77 , plus the sum of the rates for each process which is now affected by the other . between `` collisions '' due to one process
there is a random walk due to `` collisions '' of the other process .
thus , these last two rates are @xmath78 and @xmath79 . summing up the four rates we obtain @xmath80 where @xmath81 is a proportionality constant , determined by the normalization of @xmath82 in the inertial range .
in this range equations ( 20 ) , ( 29 ) and ( 31 ) yield @xmath83 while equation ( 18 ) results in @xmath84 with @xmath85 where @xmath86 is the energy rate per unit mass supplied to the turbulence from the driving source at all wavenumbers smaller than @xmath26 . combined together , equations ( 32 ) and ( 33 ) imply that @xmath87 this should coincide with @xmath82 corresponding to the kolmogorov spectral function @xmath88 thus , we obtain @xmath89 where @xmath90 is the kolmogorov constant .
we used here the same normalization with respect to @xmath90 as was used in cm ( 1991 ) . with the definition of @xmath73 , equation ( 31 ) ,
the model equations can be solved for @xmath45 , and consequently for @xmath44 and @xmath91 .
we introduce for notational convenience ( following the convention of cgc ( 1987 ) ) the rate @xmath92 defined by @xmath93 in the model for large scale turbulence of canuto & goldman ( 1985 ) @xmath94 was identified with @xmath95 , neglecting the role of the turbulence . in cgc ( 1987 )
@xmath96 was used , neglecting the role of the source .
here , @xmath94 depends both on @xmath95 and on @xmath97 , as seen from equations ( 31 ) and ( 38 ) , so it depends both on the source and on the turbulence . at @xmath34 , @xmath98 and therefore is @xmath99 times larger than in cgc ( 1987 ) . for high values of @xmath26 ( practically few times @xmath34 ) @xmath100 .
it is of interest to note that the inverse of the timescale for two - times velocity correlation according to the dia model , indeed conforms to the present definition of @xmath94 ( see figures 3 and 4 in canuto & battaglia 1988 ) . in particular at @xmath34 it is indeed @xmath101 . from equations ( 29 ) and ( 38 ) we have @xmath102 and equations ( 20 ) and ( 38 ) lead to @xmath103 combining equations ( 39 ) , ( 40 ) and ( 9 ) yields the differential equation for @xmath92 @xmath104 with @xmath0 and @xmath73 defined as functionals of @xmath92 through equations ( 28 ) and ( 31 ) , respectively . a solution for @xmath92 will also yield these two rates , as well as the spectral function @xmath45 and the mean squared turbulent vorticity @xmath82 .
the main objective of this work is the computation of the turbulent convective flux @xmath105 here @xmath106 is the turbulent velocity in the flux direction , @xmath107 is the temperature fluctuation , and the overline denotes ensemble averaging .
the dimensionless convective flux @xmath108 is determined by equations ( 12 ) and ( 34 ) to be @xmath109 the value of @xmath110 will be determined by using the fact that @xmath111 increases with @xmath26 and reaches its asymptotic value already in the inertial range .
thus , we shall follow the solution of @xmath92 and the corresponding @xmath111 , from @xmath34 up to to the inertial range , until @xmath111 saturates to @xmath110 . from equations ( 34 ) , ( 38 ) and ( 40 ) we find @xmath112 thus , once @xmath92 is obtained , @xmath111 is determined too .
turn now to the equation ( 41 ) which is a first order differential equation for @xmath92 and thus a boundary value is required for a unique solution .
below we find the value of @xmath113 .
the value of @xmath34 is determined by the width of the layer @xmath114 , @xmath115 with @xmath116 . since by definition @xmath117 and @xmath118 , equations ( 39 ) , and ( 40 )
imply that @xmath119 and @xmath120 the last two equations together with equation ( 28 ) determine @xmath113 @xmath121 where @xmath122 is the solution of @xmath123 with @xmath124 where @xmath5 is the dimensionless product of the rayleigh and prandtl numbers @xmath125 @xmath5 , or more precisely @xmath126 , is a measure of the convective efficiency . with the determination of @xmath113 , equation ( 41 ) can be solved numerically to yield @xmath92 for any value of @xmath5 .
since we deal with stellar interiors for which @xmath127 , we take in all the equations the limit of @xmath128 .
physically this means that the wavenumber at which the rate of dissipation by the microscopic viscosity becomes equal to @xmath129 , is much larger than @xmath34 , so that the spectrum exhibits an inertial range over many decades of @xmath26 . for the limit of @xmath128 , appropriate for stellar interiors , equation ( 49 )
yields an analytic solution for @xmath122 as function of @xmath5 , @xmath130 implying that @xmath131 with @xmath132 and @xmath133 .
it is convenient to express the equations in terms of normalized rates @xmath134 @xmath135 and @xmath136 and to introduce a dimensionless wavenumber @xmath137 using equations ( 26 ) , ( 27 ) and ( 28 ) in equation ( 47 ) results in @xmath138 we adopt , following cgc ( 1987 ) , @xmath139 for which equation ( 60 ) yields @xmath140 thus @xmath141 and @xmath142 the differential equation for @xmath143 is obtained from equation ( 41 ) @xmath144 where now the prime denotes differentiation with respect to @xmath145 . by definition
[ see eqns .
( 56 ) , ( 57 ) , ( 58 ) ] @xmath146 for any @xmath5 .
we have now @xmath147 where @xmath148 and @xmath149 the use of the normalized rates ensures that the computed quantities will not be very small ( large ) even for very small ( large ) @xmath5 values , thus improving the numerical accuracy . equation ( 44 ) becomes now , @xmath150 the dimensionless convective flux is given by @xmath151 where @xmath152 is the upper value of q for which equation ( 65 ) is solved , and is well inside the inertial range ( thus , @xmath153 ) .
we solved equation ( 65 ) for @xmath143 , with the initial condition of equation ( 66 ) , for @xmath5 in the range @xmath6@xmath154 .
we adopted the commonly used value of @xmath155 .
each solution yields @xmath156 and @xmath157 and the spectral function @xmath158 . for each such solution we computed @xmath159 , followed it to saturation , and thus obtained @xmath160 and @xmath161 . in figure 1
we show @xmath159 in units of @xmath162 , for @xmath163 .
the qualitative behavior of @xmath159 is typical for any value of @xmath5 : it starts from zero and saturates in the inertial range to @xmath110 . from equations ( 70 ) and ( 71 ) it follows that the asymptotic value of the graph equals @xmath108 in units of @xmath164 . in table 1
we list @xmath161 ( rounded to 4 figures ) for @xmath165 representative values of @xmath5 . from equation ( 71 )
it follows that @xmath166 .
the values shown in table 1 are in units of @xmath167 .
the limiting behavior of @xmath161 is given by @xmath168 and @xmath169 rrrrrrr @xmath6&@xmath170&@xmath171&@xmath172 & @xmath173 & @xmath174&@xmath175@xmath176&@xmath177&@xmath178&@xmath172 & @xmath179 & @xmath180&@xmath175@xmath181&@xmath182&@xmath183&@xmath184 & @xmath185 & @xmath186 & @xmath175 @xmath187&@xmath188&@xmath189&@xmath190 & @xmath191 & @xmath192 & @xmath175@xmath193&@xmath194&@xmath195&@xmath196 & @xmath197 & @xmath198 & @xmath175@xmath199&@xmath200&@xmath201&@xmath202 & @xmath203&@xmath204 & @xmath205@xmath206&@xmath207&@xmath208&@xmath209 & @xmath210 & @xmath211 & @xmath212@xmath213&@xmath214&@xmath215&@xmath216 & @xmath217 & @xmath218 & @xmath219 @xmath220&@xmath221&@xmath222&@xmath223 & @xmath224 & @xmath225 & @xmath226 @xmath227&@xmath228&@xmath229&@xmath230 & @xmath231 & @xmath232 & @xmath233@xmath234&@xmath235&@xmath236&@xmath237 & @xmath238 & @xmath239&@xmath240 @xmath241&@xmath242&@xmath243&@xmath244 & @xmath245 & @xmath246 & @xmath247 @xmath248&@xmath249&@xmath250&@xmath251 & @xmath252 & @xmath253 & @xmath254@xmath255&@xmath256&@xmath257&@xmath258 & @xmath259 & @xmath260 & @xmath261 @xmath262&@xmath263&@xmath264&@xmath265 & @xmath266 & @xmath267 & @xmath268@xmath269&@xmath270&@xmath271&@xmath272 & @xmath273 & @xmath274 & @xmath268@xmath275&@xmath276&@xmath277&@xmath278 & @xmath279 & @xmath280 & @xmath268@xmath281&@xmath282&@xmath283&@xmath284 & @xmath285 & @xmath286 & @xmath268@xmath287&@xmath288&@xmath283&@xmath289 & @xmath290 & @xmath291 & @xmath268@xmath154&@xmath292&@xmath283&@xmath293 & @xmath294 & @xmath295 & @xmath268 in applications of @xmath161 to stellar structure codes , it is useful to have an analytic fit formula to the convective flux .
we derived such a fit with a deviation @xmath296 : @xmath297 where @xmath298 where @xmath299 and @xmath300 where @xmath301 @xmath302 in order to judge the quality of the fit , in figure 2 we display the ratio between the fit function and an interpolation of the numerical values , as a function of log @xmath5 .
also shown in table 1 are the ratios @xmath303 and @xmath304 where @xmath305 and @xmath306 are the values corresponding to the mlt , @xmath307 and to the cm model ( their eq .
[ 32 ] ) . figures 3 and 4 display @xmath303 and @xmath304 , respectively . note that the qualitative behavior of @xmath303 is similar to that of the cm model higher flux for high @xmath5 and lower flux for low @xmath5 values .
the comparison with @xmath308 shows that while the two models yield essentially the same flux for high @xmath5 values , the new model predicts higher fluxes for intermediate and low @xmath5 values , and the flux ratio is maximal for @xmath309 .
comparisons of @xmath161 computed within the new model with @xmath161 computed in a model with the same definition of @xmath94 but with @xmath95 equal to the linear growth rate , indicates that the above local maximum at @xmath309 is a feature resulting from the use of the selfconsistent rate .
we recall that @xmath310 rather than @xmath5 is the measure of the convective efficiencies .
thus , the borderline between low and high efficiencies is around @xmath311 , which is also where the ratio of the new convective flux to that of the cm model is maximal .
the mean squared turbulent velocity is defined by @xmath312 which with the use of equation ( 39 ) can be expressed as @xmath313 with @xmath314 we computed @xmath315 for each value of @xmath5 and the results are presented in table 1 .
the limiting behavior is given by @xmath316 and @xmath317 the turbulent pressure is of importance in helioseismological models .
batchelor ( 1953 ) derived an expression for the mean squared turbulent pressure ( for the case of isotropic turbulence ) in terms of the spectral function @xmath45 @xmath318 where @xmath319 is the mean density and the dimensionless integral @xmath320 is @xmath321 and is maximal at @xmath322 , where it equals @xmath323 .
thus , the pressure is mostly contributed by @xmath324 which are close to the maximum of @xmath45 . using equation ( 39 )
we obtain @xmath325 the computed values of the root mean squared turbulent pressure are displayed in table 1 .
the asymptotic behavior is given by @xmath326 and @xmath327 a quantity of interest is @xmath328 defined as @xmath329 the computed values of @xmath330 are listed in table 1 . as can be seen , it is almost constant for all values of @xmath5 , ranging from @xmath331 for low @xmath5 to @xmath332 for high values of @xmath5 . as with the convective flux it is useful to have analytical fit formulae for the mean squared turbulent velocity and for the root mean squared turbulent pressure .
we derived such fits which represent the numerical values with precision better than @xmath333 .
for the mean squared velocity we derive the fit @xmath334 where @xmath335 and @xmath336 similarly , for the root mean square of the turbulent pressure we find @xmath337 with @xmath338 given by equation ( 90 ) and @xmath339 finally , the turbulent viscosity , is given in the present model already in an _
analytic _ form . from equations ( 38 ) , ( 44 ) ,
( 45 ) , ( 54 ) and ( 55 ) we find that @xmath340 the limiting behavior of @xmath67 is given by @xmath341 and @xmath342
the new convective fluxes have been included in the aton stellar structure code ( for an update on the physical and numerical details of the code , see mazzitelli et al .
1995 and references therein ) .
we have computed the main sequence evolution of a solar model as well as a set of evolutions for pop ii stars having @xmath343 , from the zero age main sequence to the base of the red giant phase . before turning to a detailed discussion of the results ,
we recall that , the turbulent length - scale @xmath114 at a given depth @xmath2 inside a convective region , must also include the thickness @xmath344 of the overshooting layers ( if any ) beyond the formal schwarzschild boundary ( see dantona and mazzitelli 1994 ) . at present
, the @xmath344 phenomenon has not yet been fully quantified in a reliable way ( umezu 1995 ) even though the underlying equations have been derived ( canuto 1993 ) .
however , empirical evidence from comparisons between stellar models constructed with local convection theories , and observations of intermediate mass main sequence stars in young open clusters , suggests quite stringent limits on the extent of the @xmath344 , namely , @xmath345 ( stothers & chin 1992 ) . lacking a formal theory , we shall write @xmath346 where @xmath347 is the pressure scale height at the upper boundary of the convective layer determined by the schwarzschild criterion and @xmath348 , which should not exceed @xmath349 , can be regarded as a _ fine - tuning _ parameter .
we stress that the role of @xmath348 in the cm and in the present models is _ radically _ different from the role of the parameter @xmath3 in the mlt model where @xmath350 .
in fact , @xmath3 is a free adjustable parameter through which modelists try to capture all the physical uncertainties ( e. g. opacities , convection , thermodynamics etc . ) relevant to the evaluation of the effective temperatures of stars .
unfortunately , not only does this procedure seriously hinder the predictive power of stellar modeling ( @xmath3 is determined _ a posteriori _ ) , but the fit to the observed surface temperature does not automatically guarantee that also the _ internal temperature profile _ is correct . in this context , it is worth noting that baturin & mironova ( 1995 ) and monteiro et al .
( 1995 ) have recently shown that the solar internal temperature profiles predicted by the cm model are in better agreement with helio - seismological data than those derived from the mlt .
moreover , gabriel ( 1995 ) has shown that the mlt could be made to predict a cm - like internal temperature profile , provided that @xmath3 is _ forced to vary inside the convective region _ , in a manner that represents an a posteriori fitting .
this quite clearly shows that a ) the mlt has no predictive power and b ) the degree of artificiality that is required to make the mlt yield results that the cm model produces quite naturally .
the parameter @xmath348 , on the other hand , quantifies a well identified _ physical process _
, the convective overshooting @xmath344 and , as seen in the following , in the present model , only a marginal amount of tuning is allowed anyhow . finally , the convective flux in table 1 is normalized to a value of the kolmogorov constant of @xmath351 , as in the cm model . since recent experimental data suggest higher values of @xmath90 ( up to @xmath352 ) , we employed for the stellar modeling a fiducial value of @xmath353 in the fit - formula , equation ( 75 ) . since , as discussed in 4.1 , the convective flux scales as @xmath354 , the flux used is a factor of @xmath355 larger than the numerical values in table 1 . in mazzitelli
et al . ( 1995 ) @xmath356 was used .
using the aton code , and updating the low
t opacities according to alexander and ferguson ( 1994 ) , we obtained a fit to the observed solar radius and luminosity at an age @xmath357 gyr ( bahcall 1989 ) , and with a metal abundance z=0.0175 ( grevesse and noels 1993 ) , with [email protected] and @xmath359 .
the latter corresponds to a very small amount of overshooting of a _ few kilometers _ , and is in full accordance with the observational limits of stothers & chin ( 1992 ) . on the other hand , had we employed the original cm model the required value would have been @xmath360 which is a borderline value . to find the maximal variance in @xmath361 allowed by the @xmath348 parameter , we computed two solar models with the borderline values @xmath362 and @xmath363 . the difference in @xmath361 between these last two models turned out to be @xmath3644% .
figure 5 shows the internal profiles of the dimensionless temperature gradient , @xmath365 , in the region of the overadiabaticity peak for solar models computed with the present , the cm and mlt models ( for the mlt model @xmath366 ) .
the results of the present model are very similar to those of the cm model but are quite different from the mlt results .
the similarity of the results of the present and cm models , in spite of the difference ( factor @xmath367 ) in the value of the convective flux for intermediate and low convective efficiencies , can be readily understood . in the more external convective layers
, the density is quite low ( @xmath368 ) , and @xmath369 .
the turbulent flux is also quite low , and convective energy transfer is very inefficient .
most of the flux in this region is therefore carried by radiation , and the temperature gradient sticks to the radiative one . in deeper convective layers , where @xmath319 , @xmath5 and @xmath114 are larger
, convection begins to be efficient and the value of the temperature gradient is determined by the turbulent convective flux .
however , for @xmath5 values such that log @xmath370 , the new fluxes are very close to the cm ones , and so is the resulting temperature gradient .
the vicinity of the gradient peak , @xmath371 , is close to the @xmath5 value where the present and cm models differ the most ( see fig .
, this is the region where we can expect some sizable difference between the temperature profiles , as figure 5 indeed shows . because of the similarity of the results from the present and the cm models , the experimental benchmark provided by helioseismological data can not discriminate between the two models ( antia and basu 1995 ) .
however , on the basis of stellar modeling , we stress that the new fluxes require a lower value of @xmath348 to fit the sun , which is more in agreement with the results of stothers & chin . since the low
t radiative opacities are probably still slightly underestimated , and larger opacities require a larger @xmath348 to fit the sun , we prefer the present new fluxes over the original cm values .
the reason is that the latter , once the updated values of low t opacities become available , could require values of @xmath348 larger than allowed by the observational upper limit on overshooting . as a further check ,
we have also applied the new fluxes to the computation of evolutionary tracks and isochrones for stars with @xmath372 and @xmath373 .
the isochrones are shown in figure 6 , together with the fiducial sequence of the globular cluster m68 .
details on the computations of both tracks and isochrones , on the observational to theoretical correlations , as well as the chemistry , reddening and distance modulus for m68 can be found in mazzitelli et al .
( 1995 ) . here
we simply recall that the apparent `` kinks '' in the isochrones are a true physical feature , which is expected and explained within the cm framework , and which exists also in the present model .
the age of the cluster is in the range @xmath374 gyr , somewhat younger than the @xmath375 gyr found with the cm fluxes , which itself is younger than the @xmath376 gyr derived within the mlt . whether this difference is significant for solving the age conflict of globular clusters with the age of the universe , following from the recent determinations of high values of @xmath377 ( freedman et al .
1994 , pierce et al .
1994 ) , is beyond the purposes of the present paper .
this point will be discussed elsewhere ( canuto et al . 1996 ) . in conclusion ,
the application of the convective fluxes of the present model to stellar structure does not alter appreciably the main results of the cm model .
nevertheless , the new model is preferable on two grounds . from the theoretical viewpoint
, it determines the rate of input of energy from buoyancy to the turbulence in a more physically consistent manner . from the astrophysical point of view , it requires a smaller extent of overshooting , which is in better agreement with recent observational results .
we have presented a selfconsistent model for turbulent convection based on a simplified treatment of the non - linear interactions among the eddies .
the important novel feature of the present model is the formulation of a selfconsistent rate for energy input from the _ source _ ( buoyancy ) into the turbulence , which depends both on the source parameters and on the _ turbulence _ itself .
this represents an improvement compared to the cm model where the rate of energy input was the growth rate of the _ linear _ unstable modes .
the focus of the present model is on the selfconsistent rate of energy input at the expense of a less complete treatment of the non - linear eddy interactions .
the latter is much simpler than in the cm model and describes transfer only from small to large wavenumbers .
this representation neglects non - local ( in @xmath26 space ) interactions between the eddies that lead to a reverse transfer ( backscatter ) that are included in the cm model .
we have explored the model for a wide range of convective efficiencies and computed , numerically , the dimensionless convective flux , the turbulent squared velocity and the root mean squared turbulent pressure , as functions of the convective efficiency @xmath5 .
the results were fitted by analytical formulae with precision better than @xmath333 over the range @xmath378 . the turbulent viscosity in the model is already given by an analytic expression .
the convective flux , is larger than that of the mlt for high convective efficiencies and lower than it for low convective efficiencies .
this general behavior is similar to that of the cm model .
the high @xmath5 fluxes are very close to those predicted by the cm model but the intermediate and lower @xmath5 fluxes are larger than those of the cm model .
it is of interest to note that even the very simple cgc ( 1987 ) model shares the same qualitative behaviour of the convective flux relative to that of the mlt flux , as function of the convective efficiency .
the fact that three models differing in their treatment of the energy input rate and in the modeling of the nonlinear transfer , still yield the same behaviour is quite intriguing .
it suggests that accounting for the full turbulence spectrum ( common to all three models ) is far more important than the detailed way in which thye latter is done .
we have applied the new model to the main sequence evolution of a solar model as well as to evolutions of pop ii stars with @xmath7 @xmath8 .
the convective turbulent length scale was taken equal to the depth in the convective zone , as in the cm model .
the results are generally similar to those of the cm model .
however , the new model has the advantage that the overshooting required to fit the solar model is smaller than in the cm model .
also , the ages of globular clusters are smaller than the corresponding ages in the cm model by @xmath379 gyr , which may help alleviate a possible conflict between the ages of globular clusters and a high value of @xmath377 .
as already noted in 5 , the similarity between the temperature profiles for the solar model , predicted by the present and the cm models , renders the models practically indistinguishable by helioseismological data ( antia & basu 1995 ) .
the situation is similar with regard to solar atmosphere modeling ( kupka 1995 ) .
however , atmospheres of cool stars are expected to yield observable differences between the two models ( kupka 1995 ) . from the theoretical perspective ,
a more complete model is one that incorporates a selfconsistent rate of energy input while keeping the non - linear interactions in their full generality , as done within the cm model .
work in this direction is in progress .
finally , the present and the cm models are based on two - point correlations of the turbulent quantities .
this methodology , preferred by the physics community , yields information about the spectral properties of the turbulence . yet
, its applicability to inhomogeneous and anisotropic cases is limited .
an alternative approach , based on one - point correlation functions , is widely used in the engineering community and can handle anisotropy and inhomogeneity .
while the spectral information is lost in this reynolds stress formalism , the method is easy to apply to non - local and space - dependent problems and has the potential to treat stellar convective overshooting ( canuto 1993 ) .
the method has already been successfully applied to the planetary boundary layer ( pbl ) which is the seat of strong convection ( canuto et al .
1994a ) as well as to study the interaction between shear , vorticity and buoyancy at the surface of the sun ( canuto et al .
work is in progress ( gabriel 1996 , houdeck 1996 ) to apply the same method to the study of helioseismology .
alexander , d.r . & ferguson , j. w , 1994 , , 437 , 879 althaus , l.g . &
benvenuto , o. g. 1996 , , 278 , 981 antia , h.m . & basu , s. 1995 , private communication batchelor , g. k. 1953 , theory of homogeneous turbulence ( cambridge : cambridge univ .
press ) bahcall , j.n .
1989 , neutrino astrophysics ( cambridge : cambridge univ . press ) , 79 baturin , v. a. & mironova , i. v. 1995 , 72 , 120 canuto , v. m. 1993 , , 416 , 331 canuto , v. m. & goldman , i. 1985 , , 54 , 430 canuto , v. m. , goldman , i. , & chasnov , j. 1987 , phys .
fluids , 30 , 3391 canuto , v. m. & battaglia , a. 1988 , , 193 , 313 canuto , v. m. & mazzitelli , i. 1991 , , 370 , 295 canuto , v. m. , cheng , y. , hartke , g. j. , & schilling , o. 1991 , phys . fluids
a , 3 , 1633 canuto , v. m. , minotti , f. o. , ronchi , c. , ypma , r. m. & zeman , o. , 1994a , j. atmos .
, 51 , 1605 canuto , v.m . , minotti , f. o. & schilling , o. 1994b , , 425 , 303 canuto , v. m. , caloi , v. , dantona f. , & mazzitelli i. 1996 , submitted gabriel , m. 1995 , a&a , 302 , 271 gabriel , m. 1996 , private communication grevesse , n. , & noels , a. 1993 , in origin and evolution of the elements , eds .
n. prantzos , e. vangioni - flam , m. cass ( cambridge : cambridge univ . press ) , 14 hartke , g. j. , canuto , v. m. , & dannevik , w. p. 1988
fluids , 31 , 256 | we present a selfconsistent model for stellar turbulent convection which is similar in spirit to the cm model ( canuto & mazzitelli 1991 ) since it accounts for the full spectrum of the turbulent eddies rather than only one eddy , as done in the mixing length theory ( mlt ) .
the model differs from the cm model in the treatment of the rate of energy input @xmath0 from the source that generates the turbulence . in the present model , @xmath0
is controlled by both the _ source _ and _ the turbulence _ it ultimately generates , thus ensuring a selfconsistent modeling of the turbulence .
this improves the cm model in which @xmath0 was taken to be equal to the growth rate of the _ linear _ unstable convective modes .
however , since the formulation of a selfconsistent treatment is far from simple , we were forced to use a representation of the nonlinear interactions less complete than the one in the cm model .
the ensuing equations were solved numerically for a wide range of convective efficiencies .
the results are the convective flux , the mean square turbulent velocity , the root mean squared turbulent pressure and the turbulent viscosity . we implemented the model in the aton stellar structure code and computed the evolution of a solar model .
the results are generally similar to those of the cm model and thus quite different from the mlt .
the present model requires a smaller overshoot into the upper radiative zone than does the cm model , in accord with recent empirical estimates .
application to pop ii stars and comparison with the very metal - poor globular cluster m68 yields an age in the range @xmath1 gyr .
this is somewhat younger than the cm age , which in turn is younger than the corresponding mlt age , a result of possible cosmological interest . # |
in this contribution we discuss signals for right - handed currents in rare hadronic @xmath0 decays .
signals in radiative @xmath0 decays are reviewed elsewhere in this report .
implications of right - handed currents for @xmath8-violation phenomenology are addressed in @xmath9 symmetric models , and in the more general case of no left - right symmetry .
we will see that it may be possible to distinguish between these scenarios at a high luminosity @xmath0 factory .
remarkably , the existence of @xmath10 symmetry could be inferred even if it is broken at a scale many orders of magnitude larger than the weak scale , e.g. , @xmath11 , in parity symmetric models @xcite .
an explicit supersymmetric realization is briefly described .
a direct test for right - handed currents from polarization measurements in @xmath0 decays to light vector meson pairs is also discussed @xcite . in the event that non - standard model @xmath8-violation is confirmed , e.g. , in the @xmath12 time - dependent @xmath8 asymmetry , an important question will be whether it arises via new physics contributions to the four - quark operators , the @xmath13 dipole operators , or both .
we will see that this question can be addressed by comparing @xmath8 asymmetries in the different transversity final states in pure penguin @xmath5 decays , e.g. , @xmath14 .
the underlying reason is large suppression of the _ transverse _ dipole operator matrix elements .
it is well known that it is difficult to obtain new @xmath15 @xmath8 violation effects at the _ loop - level _ from the _ dimension - six _ four - quark operators .
thus , this information could help discriminate between scenarios in which new physics effects are induced via loops versus at tree - level .
extensions of the standard model often include new @xmath16 right - handed currents .
these are conventionally associated with opposite chirality effective operators @xmath17 which are related to the standard model operators @xmath18 by parity transformations , * qcd penguin operators @xmath19 * chromo / electromagnetic dipole operators @xmath20 * electroweak penguin operators @xmath21 examples of new physics which could give rise to right - handed currents include supersymmetric loops which contribute to the qcd penguin or chromomagnetic dipole operators .
these are discussed in detail elsewhere in this report .
figure [ fig : susyloops ] illustrates the well known squark - gluino loops in the squark mass - insertion approximation . for example , the down - squark mass - insertion @xmath22 ( @xmath23 ) would contribute to @xmath24 ( @xmath25 ) , whereas @xmath26 ( @xmath27 ) would contribute to @xmath28 ( @xmath29 ) .
right - handed currents could also arise at tree - level via new contributions to the qcd or electroweak penguin operators , e.g. , due to flavor - changing @xmath30 couplings , @xmath31-parity violating couplings , or color - octet exchange .
we exploit the large collection of _ pure - penguin _ @xmath32 decay modes , which in the standard model have * null decay rate @xmath8-asymmetries , @xmath33 , or * null deviations of the time - dependent @xmath8-asymmetry coefficient @xmath34 from @xmath35 in decays to @xmath8-eigenstates , @xmath36 , or * null triple - product @xmath8-asymmetries @xmath37 in @xmath5 decays .
we recall that there are three helicity amplitudes @xmath38 ( @xmath39 ) in @xmath40 decays : @xmath41 , in which both vectors are longitudinaly polarized ; @xmath42 , in which both vectors have negative helicity ; and @xmath43 , in which both vectors have positive helicity . in the transversity basis @xcite ,
the amplitudes are given by , @xmath44 in @xmath0 decays , @xmath45 .
the cp - violating triple - products @xcite ( related to @xmath46 ) are then given by @xmath47 the triple - products are discussed in detail in the contribution of a. datta .
a partial list of null standard model @xmath8 asymmetries in pure - penguin decays is given below @xcite , * @xmath48 , @xmath49 , @xmath50 , @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 , ... * @xmath56 , @xmath57 , @xmath58 , @xmath59 , @xmath60 , ... * @xmath61 , @xmath62 , ... in addition , there are several modes which are penguin - dominated and are predicted to have approximately null or small standard model asymmetries , e.g. , @xmath63 ( @xmath64 subtracted ) @xcite , @xmath65 @xcite , and @xmath66 .
under parity , the effective operators transform as @xmath67 .
the new physics amplitudes , for final states @xmath68 with parity @xmath69 , therefore satisfy @xmath70 where @xmath71 and @xmath72 are the new wilson coefficient contributions to the ith pair of standard model and opposite chirality operators @xcite .
it follows that for decays to @xmath73 , @xmath74 , and @xmath75 final states , where @xmath76 , @xmath77 and @xmath78 are scalar , pseudoscalar , and vector mesons , respectively , the new physics amplitudes satisfy @xmath79 @xmath80 in @xmath5 decays the @xmath81 transversity and @xmath82 transversity final states are @xmath77-odd and @xmath77-even , respectively , yielding @xmath83 similarly , replacing one of the vector mesons with an axial - vector meson gives @xmath84 it is useful to classify the null and approximately null standard model @xmath8 asymmetries listed above according to whether the final state is @xmath77-odd or @xmath77-even , * p - even : @xmath48 , @xmath49 , @xmath85 , @xmath57 , @xmath86 , @xmath87 , @xmath53 , @xmath54 , @xmath88 , ... * p - odd : @xmath89 , @xmath56 , @xmath51 , @xmath90 , @xmath91 , @xmath92 , ... * modes with small standard model asymmetries : @xmath63 ( approximately @xmath77-even ) , @xmath65 ( p - even ) , and @xmath66 ( p - odd ) .
we are now ready to discuss implications for @xmath8 violation phenomenology in the two classes of models mentioned earlier . in the limit in which new physics is parity symmetric at the weak scale
the relation @xmath93 would hold . in light of ( [ anpparity ] ) this would imply @xcite * preservation of null @xmath8 asymmetry predictions in _ p - even _ final states . similarly , the @xmath94 constraint would be trivially satisfied .
* possibly large departures from null @xmath8 asymmetries in _
p - odd _ final states .
for example , no deviations in @xmath95 , @xmath96 , @xmath97 , @xmath98 could be accompanied by significant deviations in @xmath99 , @xmath100 , @xmath101 , and @xmath102 . both of the triple - products @xmath103 and @xmath104 in ( [ at ] ) could be affected through a modification of @xmath105 .
however , there would be no novel @xmath8 asymmetry in the interference of the parallel and longitudinal polarizations .
equivalently , the measurable quantities @xmath106 and @xmath107 defined below @xmath108 would be equal .
parity - symmetric new physics requires @xmath109 symmetry at high energies .
thus , exact weak scale parity can not be realized due to renormalization group effects below the @xmath3 breaking scale , @xmath110 .
potentially , the largest source of parity violation is the difference between the top and bottom quark yukawa couplings . in particular ,
when @xmath111 the charged higgs yukawa couplings break parity .
two scenarios for the yukawa couplings naturally present themselves : * moderate @xmath112 , or @xmath113 * maximal - parity : @xmath114 or @xmath115 small corrections to the limit of equal up and down yukawa matrices are required in order to generate the observed ckm quark mixings and light quark masses .
@xmath116 therefore sets the scale for minimal parity violation in the yukawa sector . a large hierarchy between the @xmath3 breaking scale and the weak scale can be realized naturally in supersymmetric left - right symmetric models .
these models contain two higgs bidoublet superfields @xmath117 ( or four @xmath118 doublets ) . via the ` doublet - doublet splitting ' mechanism @xcite two linear combinations of the higgs doublets
acquire masses of order @xmath110 , leaving the two light higgs doublets of the mssm .
realization of approximately parity symmetric contributions to the _ dipole _
operators favors explicit @xmath8 violation .
spontaneous @xmath8 violation could lead to complex @xmath77-violating vacuum expectation values which would feed into new loop contributions to the operators . for example , @xmath77 invariance above the weak scale would imply @xmath119 where @xmath120 breaks @xmath118 and @xmath121 is in general complex due to explicit @xmath8 violating phases .
( @xmath122 is an order tev new physics scale , e.g. , the squark or gluino masses in figure 1 ) .
thus , @xmath123 would have to be real to good approximation in order to obtain @xmath124 .
note that this also requires real gaugino masses ; otherwise rge effects would induce a complex higgs bilinear @xmath0 term in the scalar potential , thus leading to complex @xmath120 .
ordinary parity symmetry insures real @xmath125 and @xmath126 gaugino masses .
real @xmath127 gaugino masses naturally follow from the @xmath128 generalization of parity @xcite .
all the vevs entering new _ four - quark _ operator loops can , in principle , be parity neutral . therefore , real vevs are less crucial for obtaining approximately parity - symmetric four - quark operator contributions .
we have carried out a two - loop rge analysis for down squark - gluino loop contributions to the dipole operators .
choosing parity symmetric boundary conditions at @xmath110 , taking @xmath4 , and running to the weak scale we obtain * moderate @xmath112 , e.g. , @xmath129 : @xmath130 } { { \rm re } [ c^{\rm np}_{8g}(m_{w } ) + \tilde{c}^{\rm np}_{8g}(m_{w } ) ] } \le 10\% } , \qquad { \tiny \frac{{\rm im}[c_{8g}^{\rm np}(m_{w } ) - \tilde{c}^{\rm np}_{8g}(m_{w } ) ] } { { \rm i m } [ c^{\rm np}_{8g}(m_{w } ) + \tilde{c}^{\rm np}_{8g}(m_{w } ) ] } \le 10\% } \ ] ] * maximal parity , @xmath131 @xmath132 } { { \rm i m } [ c^{\rm np}_{8g}(m_{w } ) + \tilde{c}^{\rm np}_{8g}(m_{w } ) ] } = o ( 1\% ) } \ ] ] the above quantities give a measure of parity violation in the weak scale wilson coefficients .
thus , we see that for @xmath133 , new @xmath8 violating contributions to the low energy lagrangian could respect parity to @xmath134 .
precision @xmath8 violation measurements in @xmath0 decays which respect ( violate ) null sm predictions in @xmath77-even ( @xmath77-odd ) final states would therefore provide evidence for @xmath9 symmetry , even if @xmath3 is broken at the gut scale .
similar results are obtained for survival of parity in the four - quark operators @xcite .
any discussion of dipole operator phenomenology must consider the upper bound on the strange quark chromo - electric dipole moment @xmath136 , obtained from the upper bound on the @xmath135hg mercury edm @xcite .
correlations between @xmath136 and new @xmath8 violating contributions to @xmath137 , @xmath138 are most easily seen by writing the dipole operator effective hamiltonian in the weak interaction basis , @xmath139 @xmath140 and @xmath141 ( @xmath142 ) are the left - handed and right - handed down quark weak interaction eigenstates , respectively .
the mass eigenstates can be written as @xmath143 , where @xmath144 stands for the @xmath145 quarks , respectively , and @xmath146 .
the bound on @xmath136 is @xmath147 , with large theoretical uncertainty , where @xmath148 is the flavor diagonal strange quark dipole operator coefficient ( in the mass eigentate basis ) .
it is given as @xmath149 similarly , the @xmath13 wilson coefficients are given as @xmath150 if significant contributions to the ckm matrix elements are generated in the down quark sector , then @xmath151 , @xmath152 , and @xmath153 . in the absence of special flavor symmetries , similar magnitudes would be expected for the corresponding right - handed quark mixing coefficients , @xmath154 .
generically , we therefore expect @xmath155 .
@xmath156 would correspond to @xmath157\sim 1 $ ] .
thus , @xmath15 @xmath8 violating effects generically correspond to a value for @xmath136 which is a factor of 100 too large . one way to evade this bound is by invoking some mechanism , e.g. , flavor symmetries , for generating the large hierarchies @xmath158 and @xmath159 . an elegant alternative is provided by parity symmetry @xcite . it is well known that edm s must vanish in the parity symmetric limit , see e.g. @xcite .
for example , in ( [ css ] ) exact parity would imply @xmath160 , @xmath161 and real @xmath162 , thus yielding a real coefficient , @xmath148 .
an rge analysis along the lines discussed above is required in order to determine the extent to which this can be realized at low energies .
we find that in both the maximal parity scenario ( @xmath131 ) and in moderate @xmath112 scenarios it is possible to obtain @xmath156 and at the same time satisfy the bound on @xmath136 if @xmath163 @xcite . in the parity - symmetric scenario ,
an unambiguous theoretical interpretation of the pattern of @xmath8 violation is possible because null predictions are maintained for the @xmath77-even final states .
however , if new contributions to the @xmath18 and @xmath17 operators are unrelated , then @xmath8 asymmetries in the @xmath77-odd and @xmath77-even null standard model modes could differ significantly both _ from each other , and from the null predictions_. this is due to the opposite relative sign between the left - handed and right - handed new physics amplitudes for @xmath77-odd and @xmath77-even final states in eqs .
( [ anpparity]) ( [ anpav ] ) .
for example , @xmath56 and @xmath57 could be affected differently in the mssm @xcite .
an interesting illustration would be provided by models with @xmath164 contributions to the @xmath165 , and negligible new contributions to the @xmath18 .
this could happen , for example , in supersymmetric models with large ( negligible ) @xmath166 squark mixing @xcite@xcite , or in models in which @xmath31-parity violation induces opposite chirality four - quark operators at the tree - level @xcite .
unrelated right - handed currents could also arise in warped extra dimension models with bulk left - right symmetry @xcite .
unfortunately , @xmath8 asymmetry predictions have large theoretical uncertainties due to @xmath167 power corrections , especially from the qcd penguin annihilation amplitudes .
they are therefore difficult to interpret .
an illustration is provided in figure [ fig : cpcomparison ] , which compares predictions for @xmath168 and @xmath169 arising from new contributions to @xmath24 and @xmath170 in qcd factorization @xcite .
for @xmath99 we take @xmath171 . for @xmath172 two corresponding cases
are considered : ( a ) a purely left - handed current , @xmath173 , @xmath174 , ( b ) a purely right - handed current , @xmath175 , @xmath176 .
the scatter plots scan over the input parameter ranges given in @xcite ( with the exception of the gegenbauer moments of the light meson light - cone distribution amplitudes and @xmath177 , which have been set to their default values ) .
in addition , the branching ratios are required to lie within their 90% c.l .
intervals .
clearly , very different values for the two @xmath8 asymmetries can be realized if the new physics only appears in @xmath24 .
for example , for @xmath178 , it is possible to obtain @xmath179 and @xmath180 .
the theoretical uncertainty in @xmath95 is larger than for @xmath181 .
we therefore expect that even larger differences are possible between @xmath95 and @xmath56 , for purely left - handed currents .
however , figure [ fig : cpcomparison ] suggests that @xmath182 and @xmath183 ( @xmath184 is realized at @xmath185 ) could be a signal for right - handed currents @xcite .
more theoretical studies are needed in order to determine if this is indeed the case .
in particular , a more thorough analysis of uncertainties due to @xmath186 effects needs to be undertaken .
for example , power corrections to the dipole operator matrix elements remain to be included .
furthermore , the impact on @xmath187 of new physics in all of the ` left - handed ' four - quark operators needs to be thoroughly studied .
a discussion of polarization in @xmath5 decays has been presented in @xcite in the framework of qcd factorization . here
we summarize some of the results . to begin with we note that the polarization should be sensitive to the @xmath188 structure of the standard model , due to the power suppression associated with the ` helicity - flip ' of a collinear quark .
for example , in the standard model the factorizable graphs for @xmath189 are due to transition operators with chirality structures @xmath190 , see figure [ fig : factorizable ] . in the helicity amplitude
@xmath191 a collinear @xmath192 or @xmath193 quark with positive helicity ends up in the negatively polarized @xmath64 , whereas in @xmath194 a second quark
helicity - flip is required in the form factor transition .
collinear quark helicity flips require transverse momentum , @xmath195 , implying a suppression of @xmath196 per flip . in the case of new right - handed currents , e.g. , @xmath197 ,
the helicity amplitude hierarchy would be inverted , with @xmath194 and @xmath198 requiring one and two helicity - flips , respectively . in naive factorization
the @xmath199 helicity amplitudes , supplemented by the large energy form factor relations @xcite , satisfy @xmath200 @xmath201 and @xmath202 are the @xmath203 form factors in the large energy limit @xcite . both scale as @xmath204 in the heavy quark limit , implying @xmath205 .
@xmath206 parametrizes form factor helicity suppression .
it is given by @xmath207 where @xmath208 and @xmath78 are the axial - vector and vector current form factors , respectively .
the large energy relations imply that @xmath209 vanishes at leading power , reflecting the fact that helicity suppression is @xmath210 .
thus , @xmath211 .
light - cone qcd sum rules @xcite , and lattice form factor determinations scaled to low @xmath212 using the sum rule approach @xcite , give @xmath213 ; qcd sum rules give @xmath214 @xcite ; and the bsw model gives @xmath215 @xcite .
the polarization fractions in the transversity basis ( [ transversity ] ) therefore satisfy @xmath216 in naive factorization , where the subscript @xmath217 refers to longitudinal polarization , @xmath218 , and @xmath219 .
the measured longitudinal fractions for @xmath220 are close to 1 @xcite .
this is not the case for @xmath221 for which full angular analyses yield @xmath222 naively averaging the belle and babar measurements ( without taking correlations into account ) yields @xmath223 .
we must go beyond naive factorization in order to determine if the small value of @xmath224 could simply be due to the dominance of qcd penguin operators in @xmath225 decays , rather than new physics .
in particular , it is necessary to determine if the power counting in ( [ smpred ] ) is preserved by non - factorizable graphs , i.e. , penguin contractions , vertex corrections , spectator interactions , annihilation graphs , and graphs involving higher fock - state gluons .
this question can be addressed in qcd factorization @xcite . in qcd factorization
exclusive two - body decay amplitudes are given in terms of convolutions of hard scattering kernels with meson light - cone distribution amplitudes @xcite@xcite . at leading power
this leads to factorization of short and long - distance physics .
this separation breaks down at sub - leading powers with the appearance of logarithmic infrared divergences , e.g. , @xmath226 , where @xmath227 is the light - cone quark momentum fraction in a final state meson , and @xmath228 is a physical infrared cutoff .
nevertheless , the power - counting for all amplitudes can be obtained . the extent to which it holds numerically
can be determined by assigning large uncertainties to the logarithmic divergences .
fortunately , certain polarization observables are less sensitive to this uncertainty , particularly after experimental constraints , e.g. , total rate or total transverse rate , are imposed .
examples of logarithmically divergent hard spectator interaction and qcd penguin annihilation graphs are shown in figure [ fig : annihilation ] , with the quark helicities indicated .
the power counting for the helicity amplitudes of the annihilation graph , including logarithmic divergences , is @xmath229 the logarithmic divergences are associated with the limit in which both the @xmath192 and @xmath193 quarks originating from the gluon are soft .
the annihilation topology implies an overall factor of @xmath230 .
each remaining factor of @xmath231 is associated with a quark helicity flip .
in fact , adding up all of the helicity amplitude contributions in qcd factorization _ formally _ preserves the naive factorization power counting in ( [ smpred ] ) @xcite .
recently , the first relation in ( [ smpred ] ) has been confirmed in the soft collinear effective theory @xcite .
however , as we will see below , it need not hold numerically because of qcd penguin annihilation .
the numerical inputs are given in @xcite .
the logaritmic divergences are modeled as in @xcite .
for example , in the annihilation amplitudes the quantities @xmath232 are introduced as @xmath233 this parametrization reflects the physical @xmath234 cutoff , and allows for large strong phases @xmath235 $ ] from soft rescattering .
the quantities @xmath232 ( and the corresponding hard spectator interaction quantities @xmath236 ) are varied independently for unrelated convolution integrals .
the predicted longitudinal polarization fractions @xmath237 and @xmath238 are close to unity , in agreement with observation @xcite and with naive power counting ( [ smpred ] ) .
the theoretical uncertainties are small , particularly after imposing the branching ratio constraints , due to the absence of ( for @xmath239 ) or ckm suppression of ( for @xmath240 ) the qcd penguin amplitudes . averaging the belle and babar @xmath241 measurements @xcite yields @xmath242 and @xmath243 , or @xmath244 and @xmath245 .
@xmath246 and @xmath247 are the @xmath8-averaged longitudinal and total transverse branching ratios , respectively . in the absence of annihilation ,
the predicted branching ratios are @xmath248 and @xmath249 , where the second ( first ) set of error bars is due to variations of @xmath236 ( all other inputs ) .
however , the @xmath250 qcd penguin annihilation graph in figure [ fig : annihilation ] can play an important role in both @xmath251 and @xmath198 due to the appearance of a logarithmic divergence squared ( @xmath252 ) , the large wilson coefficient @xmath253 , and a @xmath254 rather than @xmath255 dependence .
although formally @xmath256 , see ( [ a3fpowers ] ) , these contributions can be @xmath164 numerically .
this is illustrated in figure [ fig : phikstpolarization ] , where @xmath257 and @xmath258 are plotted versus the quantities @xmath259 and @xmath260 , respectively , for @xmath261 .
@xmath259 and @xmath260 enter the parametrizations ( [ xa ] ) of the logarithmic divergences appearing in the longitudinal and negative helicity @xmath250 annihilation amplitudes , respectively .
as @xmath262 increase from 0 to 1 , the corresponding annihilation amplitudes increase by more than an order of magnitude .
the theoretical uncertainties on the rates are very large .
furthermore , the largest input parameter uncertainties in @xmath246 and @xmath258 are a priori unrelated .
thus , it is clear from figure [ fig : phikstpolarization ] that the qcd penguin annihilation amplitudes can account for the @xmath263 measurements .
similarly , the babar measurement of @xmath264 @xcite can be accounted for .
do the qcd penguin annihilation amplitudes also imply large transverse polarizations in @xmath265 decays ?
the answer depends on the pattern of @xmath266 flavor symmetry violation in these amplitudes . for light mesons containing a single strange quark ,
e.g. , @xmath267 , non - asymptotic effects shift the weighting of the meson distribution amplitudes towards larger strange quark momenta . as a result ,
the suppression of @xmath268 popping relative to light quark popping in annihilation amplitudes can be @xmath164 , which is consistent with the order of magnitude hierarchy between the @xmath269 and @xmath270 rates @xcite .
( see @xcite for a discussion of other sources of @xmath271 violation ) . in the present case , this implies that the longitudinal polarizations should satisfy @xmath272 in the standard model @xcite .
consequently , @xmath273 would suggest that @xmath274-spin violating new physics entering mainly in the @xmath275 channel is responsible for the small @xmath276 .
one possibility would be right - handed vector currents ; they could interfere constructively ( destructively ) in @xmath277 ( @xmath278 ) transversity amplitudes , see ( [ anpvv ] ) .
alternatively , a parity - symmetric scenario would only affect @xmath277 .
a more exotic possibility would be tensor currents ; they would contribute to the longitudinal and transverse amplitudes at sub - leading and leading power , respectively , opposite to the vector currents .
we should mention that our treatment of the charm ( and up ) quark loops in the penguin amplitudes follows the usual perturbative approach used in qcd factorization @xcite@xcite .
the authors of @xcite believe that the region of phase space in which the charm quark pair has invariant mass @xmath279 , and is thus moving non - relativistically , should be separated out into a long - distance ` charming penguin ' amplitude @xcite .
nrqcd arguments are invoked to claim that such contributions are @xmath280 , where @xmath281 , so that they could effectively be of leading power .
furthermore , it is claimed that the transverse components may also be of leading power , thus potentially accounting for @xmath224 . however , a physical mechanism by which a collinear quark helicity - flip could arise in this case without power suppression remains to be clarified .
arguments against a special treatment of this region of phase space @xcite are based on parton - hadron duality .
it should be noted that in qcd factorization this region of @xmath212 contributes negligibly to the @xmath5 penguin amplitudes , particularly in the transverse components .
more recently , the low value of @xmath224 has been addressed using a purely hadronic model for soft rescattering of intermediate two - body charm states , i.e. , @xmath282 @xcite .
this approach has been criticized previously on the grounds that a `` purely hadronic language , suitable for kaon decays '' is not applicable to the case of @xmath0 decays , where the `` number of channels , and the energy release are large '' @xcite . in particular , many intermediate multi - body channels have been ignored which are predicted to lead to systematic amplitude cancelations in the heavy quark limit .
does the naive factorization relation @xmath283 ( [ smpred ] ) survive in qcd factorization ?
this ratio is very sensitive to the quantity @xmath284 defined in ( [ rperp ] ) . as @xmath206 increases
, @xmath285 decreases .
the range @xmath286 spanning existing model determinations @xcite@xcite is taken in @xcite . in figure
[ fig : fperppar ] ( left ) the resulting predictions for @xmath287 and @xmath258 are studied simultaneaously for @xmath261 in the standard model .
note that the theoretical uncertainty for @xmath287 is much smaller than for @xmath288 .
evidently , the above relation still holds , particularly at larger values of @xmath258 where qcd penguin annihilation dominates both @xmath289 and @xmath290 . a ratio for @xmath287 in excess of the standard model range , e.g. , @xmath291 if @xmath292 , would signal the presence of new right - handed currents .
this is due to the inverted hierarchy between @xmath198 and @xmath194 for right - handed currents , and is reflected in the sign difference with which the wilson coefficients @xmath293 enter @xmath294 and @xmath295 . for illustration , new contributions to the qcd penguin operators are considered in figure [ fig : fperppar ] ( right ) . at the new physics matching scale @xmath296 , these can be parametrized as @xmath297 . for simplicity , we take @xmath298 and consider two cases : @xmath299 or new left - handed currents ( lower bands ) , and @xmath300 or new right - handed currents ( upper bands ) , corresponding to @xmath301 or @xmath302 , and @xmath303 or @xmath304 . clearly
, moderately sized right - handed currents could increase @xmath285 well beyond the standard model range if @xmath305 . however , new left - handed currents would have little effect .
the @xmath306 penguin contractions of the chromomagnetic dipole operator @xmath24 are illustrated in figure [ fig : q8 g ] .
@xmath307 and @xmath308 are the qcd factorization coefficients of the transition operators @xmath309 and @xmath310 , respectively , where @xmath311 is summed over @xmath312 @xcite . only the contribution on the left ( @xmath307 ) to the longitudinal helicity amplitude
@xmath251 is non - vanishing @xcite . in particular , the chromo- and electromagnetic dipole operators @xmath24 and @xmath313 _ do not contribute to the transverse penguin amplitudes _ at @xmath306 due to angular momentum conservation : the dipole tensor current couples to a transverse gluon , but a ` helicity - flip ' for @xmath311 or @xmath314 in figure [ fig : q8 g ] would require a longitudinal gluon coupling .
formally , this result follows from wandura - wilczek type relations among the vector meson distribution amplitudes , and the large energy relations between the tensor - current and vector - current form factors .
transverse amplitudes in which a vector meson contains a collinear higher fock state gluon also vanish at @xmath315 , as can be seen from the vanishing of the corresponding partonic dipole operator graphs in the same momentum configurations .
furthermore , the transverse @xmath316 contributions involving spectator interactions are highly suppressed .
this has important implications for new physics searches .
for example , in pure penguin decays to cp - conjugate final states @xmath68 , e.g. , @xmath317 , if the transversity basis time - dependent cp asymmetry parameters @xmath318 and @xmath319 are consistent with @xmath320 , and @xmath321 is not , then this would signal new cp violating contributions to the chromomagnetic dipole operators .
however , deviations in @xmath322 or @xmath323 would signal new cp violating four - quark operator contributions .
if the triple - products @xmath324 and @xmath325 ( [ at ] ) do not vanish and vanish , respectively , in pure - penguin decays , then this would also signal new cp violating contributions to the chromomagnetic dipole operators .
this assumes that a significant strong phase difference is measured between @xmath326 and @xmath327 , for which there is some experimental indication @xcite . however , non - vanishing @xmath328 , or non - vanishing transverse direct cp asymmetries would signal the intervention of four - quark operators .
the above would help to discriminate between different explanations for an anomalous @xmath99 , which fall broadly into two categories : radiatively generated dipole operators , e.g. , supersymmetric loops ; or tree - level four - quark operators , e.g. , flavor changing ( leptophobic ) @xmath329 exchange @xcite , @xmath31-parity violating couplings @xcite , or color - octet exchange @xcite . finally , a large value for @xmath287 would be a signal for right - handed _ four - quark _ operators .
there are a large number of penguin - dominated rare hadronic @xmath0 decay modes in the standard model in which departures from null @xmath8 asymmetry predictions would be a signal for new physics .
we have seen that in order to detect the possible intervention of new @xmath16 right - handed currents it is useful to organize these modes according to the parity of the final state .
@xmath9 symmetric models in which new @xmath8 violating contributions to the effective @xmath330 hamiltonian are to good approximation parity symmetric at the weak scale , would only give rise to significant deviations from null @xmath8 asymmetries in parity - odd final states .
for example , no deviations from the null standard model @xmath8 asymmetry predictions in @xmath95 , @xmath96 , @xmath331 , @xmath98 could be accompanied by significant deviations in @xmath99 , @xmath100 , @xmath332 @xmath101 , and @xmath102 .
this would provide a clean signal for left - right symmetry .
however , the precision of @xmath8 asymmetry measurements necessary to discern the existence of such a pattern would require a high luminosity @xmath0 factory .
remarkably , approximate parity invariance in the @xmath333 effective hamiltonian can be realized even if the @xmath3 symmetry breaking scale @xmath110 is as large as @xmath334 .
an explicit example in which large departures from the null predictions are possible , but in which deviations from parity invariance can be as small as @xmath2 for @xmath4 , is provided by squark - gluino loops in parity - symmetric susy models .
it is noteworthy that , due to parity invariance , stringent bounds on new sources of @xmath8 and flavor violation arising from the @xmath135hg mercury edm are naturally evaded in such models .
more generally , in models in which new contributions to standard model ( left - handed ) and opposite chirality ( right - handed ) effective operators are unrelated , the @xmath8 asymmetries in the @xmath77-odd and @xmath77-even null standard model modes could differ substantially both from each other , and from the null predictions .
this is because the right - handed operator wilson coefficients enter with opposite sign in the amplitudes for decays to @xmath77-odd and @xmath77-even final states .
unfortunately , @xmath8 asymmetry predictions have large theoretical uncertainties due to @xmath167 power corrections , especially from the qcd penguin annihilation amplitudes .
we therefore can not rule out substantial differences between new @xmath8 violating effects in parity - even and parity - odd modes arising solely from left - handed currents .
however , very large differences , e.g. , @xmath182 and @xmath183 , may provide a signal for @xmath8 violating right - handed currents .
more theoretical work will be required in order to make this statement more precise .
polarization measurements in @xmath0 decays to light vector meson pairs offer a unique opportunity to probe the chirality structure of rare hadronic @xmath0 decays . a standard model analysis which includes all
non - factorizable graphs in qcd factorization shows that the longitudinal polarization formally satisfies @xmath335 , as in naive factorization .
however , the contributions of a particular qcd penguin annihilation graph which is formally @xmath336 can be @xmath15 numerically in longitudinal and negative helicity @xmath337 @xmath338 decays . consequently
, the observation of @xmath339 can be accounted for , albeit with large theoretical errors .
the expected pattern of @xmath266 violation in the qcd penguin annihilation graphs , i.e. , large suppression of @xmath340 relative to @xmath341 or @xmath342 popping , implies that the longitudinal polarizations should satisfy @xmath272 in the standard model .
consequently , @xmath273 would suggest that @xmath274-spin violating new physics entering mainly in the @xmath275 channel is responsible for the small values of @xmath276 .
the ratio of transverse rates in the transversity basis satisfies @xmath343 , in agreement with naive power counting . a ratio in excess of the predicted standard model range would signal the presence of new right - handed currents in dimension-6 four - quark operators .
the maximum ratio attainable in the standard model is sensitive to the @xmath344 form factor combination @xmath206 , see ( [ rperp ] ) , which controls helicity suppression in form factor transitions .
all existing model determinations give a positive sign for @xmath206 , which would imply @xmath345 in the standard model .
the magnitude and especially the sign of @xmath346 is clearly an important issue which should be clarified further with dedicated lattice studies .
contributions of the dimension-5 @xmath13 dipole operators to the transverse @xmath5 modes are highly suppressed , due to angular momentum conservation .
comparison of @xmath8 violation involving the longitudinal modes with @xmath8 violation only involving the transverse modes in pure penguin @xmath225 decays could therefore distinguish between new contributions to the dipole and four - quark operators .
more broadly , this could distinguish between scenarios in which new physics effects are loop induced and scenarios in which they are tree - level induced , as it is difficult to obtain @xmath15 cp - violating effects from dimension-6 operators beyond tree - level .
again , a high luminosity @xmath0 factory will be required in order to obtain the necessary level of precision in @xmath8 violation measurements .
i would like to thank martin beneke , alakhabba datta , keith ellis , guy engelhard , andrei gritsan , yuval grossman , matthias neubert , uli nierste , yossi nir , dan pirjol , and ian stewart for useful discussions .
this work was supported by the department of energy under grant de - fg02 - 84er40153 .
g. valencia , phys .
d * 39 * , 3339 ( 1989 ) ; a. datta and d. london , arxiv : hep - ph/0303159 .
for early references on new physics and null sm modes see : y. grossman and m. p. worah , phys .
b * 395 * , 241 ( 1997 ) [ arxiv : hep - ph/9612269 ] ; r. fleischer , int .
j. mod .
a * 12 * , 2459 ( 1997 ) [ arxiv : hep - ph/9612446 ] , and in proceedings of 7th international symposium on heavy flavor physics , santa barbara , jul 1997 [ arxiv : hep - ph/9709291 ] ; d. london and a. soni , phys . lett .
b * 407 * , 61 ( 1997 ) [ arxiv : hep - ph/9704277 ] ; m. ciuchini , e. franco , g. martinelli , a. masiero and l. silvestrini , phys .
* 79 * , 978 ( 1997 ) [ arxiv : hep - ph/9704274 ] ; a. l. kagan , in proceedings of 2nd international conference on b physics and cp violation , honolulu , mar 1997 , and 7th international symposium on heavy flavor physics , santa barbara , jul 1997 [ arxiv : hep - ph/9806266 ] .
a. garmash _ et al . _ [ belle collaboration ] , phys .
d * 69 * , 012001 ( 2004 ) [ arxiv : hep - ex/0307082 ] ; b. aubert _ et al . _ [ babar collaboration ] , arxiv : hep - ex/0406005 .
y. grossman , z. ligeti , y. nir and h. quinn , phys .
d * 68 * , 015004 ( 2003 ) [ arxiv : hep - ph/0303171 ] ; m. gronau and j. l. rosner , phys .
b * 564 * , 90 ( 2003 ) [ arxiv : hep - ph/0304178 ] .
d. london and a. soni , in @xcite ; m. gronau , y. grossman and j. l. rosner , phys .
b * 579 * , 331 ( 2004 ) [ arxiv : hep - ph/0310020 ] .
a. l. kagan , talk at slac summer institute , august 2002 .
k. s. babu , b. dutta and r. n. mohapatra , phys .
d * 65 * , 016005 ( 2002 ) [ arxiv : hep - ph/0107100 ] .
r. n. mohapatra and a. rasin , phys .
d * 54 * , 5835 ( 1996 ) [ arxiv : hep - ph/9604445 ]
. t. falk , k. a. olive , m. pospelov and r. roiban , nucl .
b * 560 * , 3 ( 1999 ) [ arxiv : hep - ph/9904393 ] ; m. v. romalis , w. c. griffith and e. n. fortson , phys .
* 86 * , 2505 ( 2001 ) [ arxiv : hep - ex/0012001 ] ; j. hisano and y. shimizu , phys .
b * 581 * , 224 ( 2004 ) [ arxiv : hep - ph/0308255 ] .
a. l. kagan , talks at 19th international workshop on weak interactions and neutrinos , geneva wi , oct 2003 , and 2004 phenomenology symposium , madison , w may 2004 .
d. chang , a. masiero and h. murayama , phys .
d * 67 * , 075013 ( 2003 ) [ arxiv : hep - ph/0205111 ] ; r. harnik , d. t. larson , h. murayama and a. pierce , phys .
d * 69 * , 094024 ( 2004 ) [ arxiv : hep - ph/0212180 ] .
a. datta , phys .
d * 66 * , 071702 ( 2002 ) [ arxiv : hep - ph/0208016 ] ; b. dutta , c. s. kim and s.
oh , phys . rev .
lett . * 90 * ( 2003 ) 011801 [ arxiv : hep - ph/0208226 ] .
g. burdman , phys .
b * 590 * , 86 ( 2004 ) [ arxiv : hep - ph/0310144 ] .
m. beneke , g. buchalla , m. neubert and c. t. sachrajda , phys .
lett . * 83 * , 1914 ( 1999 ) [ hep - ph/9905312 ] ; nucl .
b * 591 * , 313 ( 2000 ) [ hep - ph/0006124 ] ; m. beneke , g. buchalla , m. neubert and c. t. sachrajda , nucl .
b * 606 * , 245 ( 2001 ) [ arxiv : hep - ph/0104110 ] .
m. beneke and m. neubert , nucl .
b * 675 * , 333 ( 2003 ) [ arxiv : hep - ph/0308039 ] .
j. charles , a. le yaouanc , l. oliver , o. pene and j. c. raynal , phys .
d * 60 * , 014001 ( 1999 ) [ arxiv : hep - ph/9812358 ] .
p. ball and v. m. braun , phys .
d * 58 * , 094016 ( 1998 ) [ hep - ph/9805422 ] .
l. del debbio , j. m. flynn , l. lellouch and j. nieves , phys .
b * 416 * , 392 ( 1998 ) [ hep - lat/9708008 ] .
p. colangelo , f. de fazio , p. santorelli and e. scrimieri , phys .
d * 53 * , 3672 ( 1996 ) [ erratum - ibid .
d * 57 * , 3186 ( 1998 ) ] [ arxiv : hep - ph/9510403 ] .
m. bauer , b. stech and m. wirbel , z. phys .
c * 34 * , 103 ( 1987 ) .
j. zhang _
et al . _ [ belle collaboration ] , phys . rev .
* 91 * , 221801 ( 2003 ) [ arxiv : hep - ex/0306007 ] .
b. aubert _ et al . _ [ babar collaboration ] , phys . rev .
lett . * 91 * , 171802 ( 2003 ) [ arxiv : hep - ex/0307026 ] ; b. aubert _ et al .
_ [ babar collaboration ] , arxiv : hep - ex/0404029 .
j. smith [ babar collaboration ] , talk at rencontres de moriond qcd04 , march 2004 [ arxiv : hep - ex/0406063 ] ; a. gritsan [ babar collaboration ] , lbnl seminar , april 2004 [ http://costard.lbl.gov/
gritsan / rpm / babar - coll-0028.pdf ] .
a. l. kagan , talks at super b factory workshop , honolulu , january 2004 ; les rencontres de physique de la valle daoste , february 2004 . c. w. bauer , d. pirjol , i. z. rothstein and i. w. stewart , arxiv : hep - ph/0401188 v2 .
s. mantry , d. pirjol and i. w. stewart , phys .
d * 68 * , 114009 ( 2003 ) [ arxiv : hep - ph/0306254 ] .
m. beneke , econf * c0304052 * ( 2003 ) f0001 [ arxiv : hep - ph/0308040 ] . | precision cp violation measurements in rare hadronic @xmath0 decays could provide clean signatures of parity symmetric new physics , implying the existence of @xmath1 symmetry at high energies .
new contributions to the weak scale hamiltonian which respect parity to @xmath2 in supersymmetric realizations are compatible with an @xmath3 breaking scale @xmath4 , implying that sensitivity to the gut scale is possible .
the generic case of right - handed currents without left - right symmetry is also discussed .
a detailed analysis of @xmath5 polarization in qcd factorization reveals that the low longitudinal polarization fraction @xmath6 can be accounted for in the sm via a qcd penguin annihilation graph .
the ratio of transverse rates @xmath7 provides a sensitive test for new right - handed currents .
cp violation measurements in @xmath5 decays can discriminate between new contributions to the dipole and four quark operators .
hep - ph/0407076 + july 6 , 2004 * right - handed currents , @xmath8 violation , and @xmath5 * alexander l. kagan + |
these notes summarize the lectures presented at the 2005 cern - claf school in malarge , argentina and the 2009 cern - claf school in medelln , colombia .
the audience in both occasions was composed to a large extent by students in experimental high energy physics with an important minority of theorists .
in nearly ten hours it is quite difficult to give a reasonable introduction to a subject as vast as quantum field theory .
for this reason the lectures were intended to provide a review of those parts of the subject to be used later by other lecturers .
although a cursory acquaitance with th subject of quantum field theory is helpful , the only requirement to follow the lectures it is a working knowledge of quantum mechanics and special relativity .
the guiding principle in choosing the topics presented ( apart to serve as introductions to later courses ) was to present some basic aspects of the theory that present conceptual subtleties .
those topics one often is uncomfortable with after a first introduction to the subject . among them
we have selected : * the need to introduce quantum fields , with the great complexity this implies .
* quantization of gauge theories and the rle of topology in quantum phenomena .
we have included a brief study of the aharonov - bohm effect and dirac s explanation of the quantization of the electric charge in terms of magnetic monopoles . * quantum aspects of global and gauge symmetries and their breaking . * anomalies . * the physical idea behind the process of renormalization of quantum field theories . * some more specialized topics , like the creation of particle by classical fields and the very basics of supersymmetry .
these notes have been written following closely the original presentation , with numerous clarifications .
sometimes the treatment given to some subjects has been extended , in particular the discussion of the casimir effect and particle creation by classical backgrounds .
since no group theory was assumed , we have included an appendix with a review of the basics concepts . by lack of space and purpose ,
few proofs have been included . instead
, very often we illustrate a concept or property by describing a physical situation where it arises .
full details and proofs can be found in the many textbooks in the subject , and in particular in the ones provided in the bibliography @xcite .
specially modern presentations , very much in the spirit of these lectures , can be found in references @xcite .
we should nevertheless warn the reader that we have been a bit cavalier about references .
our aim has been to provide mostly a ( not exhaustive ) list of reference for further reading .
we apologize to those authors who feel misrepresented .
* acknowlegments .
* it is a great pleasure to thank the organizers of the cern - claf 2005 and 2009 schools , and in particular teresa dova , marta losada and enrico nardi , for the opportunity to present this material , and for the wonderful atmosphere they created during the schools .
the work of m.a.v .- m . has been partially supported by spanish science ministry grants fpa2002 - 02037 , fpa2005 - 04823 and bfm2003 - 02121 . before starting it is convenient to review the notation used . through these notes
we will be using the metric @xmath0 .
derivatives with respect to the four - vector @xmath1 will be denoted by the shorthand @xmath2 as usual space - time indices will be labelled by greek letters ( @xmath3 ) while latin indices will be used for spatial directions ( @xmath4 ) . in many expressions
we will use the notation @xmath5 where @xmath6 are the pauli matrices @xmath7 sometimes we use of the feynman s slash notation @xmath8 . finally , unless stated otherwise , we work in natural units @xmath9 .
in spite of the impressive success of quantum mechanics in describing atomic physics , it was immediately clear after its formulation that its relativistic extension was not free of difficulties .
these problems were clear already to schrdinger , whose first guess for a wave equation of a free relativistic particle was the klein - gordon equation @xmath10 this equation follows directly from the relativistic `` mass - shell '' identity @xmath11 using the correspondence principle @xmath12 plane wave solutions to the wave equation ( [ kg ] ) are readily obtained @xmath13 in order to have a complete basis of functions , one must include plane wave with both @xmath14 and @xmath15 .
this implies that given the conserved current @xmath16 its time - component is @xmath17 and therefore does not define a positive - definite probability density .
a complete , properly normalized , continuous basis of solutions of the klein - gordon equation ( [ kg ] ) labelled by the momentum @xmath18 can be defined as @xmath19 given the inner product @xmath20 the states ( [ basis ] ) form an orthonormal basis @xmath21 = 0.5truein the wave functions @xmath22 describes states with momentum @xmath18 and energy given by @xmath23 . on the other hand , the states @xmath24 not only
have a negative scalar product but they actually correspond to negative energy states @xmath25 therefore the energy spectrum of the theory satisfies @xmath26 and is unbounded from below ( see fig .
[ fig : spectrum ] ) .
although in a case of a free theory the absence of a ground state is not necessarily a fatal problem , once the theory is coupled to the electromagnetic field this is the source of all kinds of disasters , since nothing can prevent the decay of any state by emission of electromagnetic radiation .
the problem of the instability of the `` first - quantized '' relativistic wave equation can be heuristically tackled in the case of spin-@xmath27 particles , described by the dirac equation @xmath28 where @xmath29 and @xmath30 are @xmath31 matrices @xmath32 with @xmath6 the pauli matrices , and the wave function
@xmath33 has four components . the wave equation ( [ eq : dirac ] ) can be thought of as a kind of `` square root '' of the klein - gordon equation ( [ kg ] ) , since the latter can be obtained as @xmath34 an analysis of eq .
( [ eq : dirac ] ) along the lines of the one presented above for the klein - gordon equation leads again to the existence of negative energy states and a spectrum unbounded from below as in fig .
[ fig : spectrum ] .
dirac , however , solved the instability problem by pointing out that now the particles are fermions and therefore they are subject to pauli s exclusion principle . hence , each state in the spectrum can be occupied by at most one particle , so the states with @xmath35 can be made stable if we assume that _ all _ the negative energy states are filled .
= 5.0truein if dirac s idea restores the stability of the spectrum by introducing a stable vacuum where all negative energy states are occupied , the so - called dirac sea , it also leads directly to the conclusion that a single - particle interpretation of the dirac equation is not possible .
indeed , a photon with enough energy ( @xmath36 ) can excite one of the electrons filling the negative energy states , leaving behind a `` hole '' in the dirac see ( see fig .
[ fig : dirac_sea ] ) .
this hole behaves as a particle with equal mass and opposite charge that is interpreted as a positron , so there is no escape to the conclusion that interactions will produce pairs particle - antiparticle out of the vacuum . in spite of the success of the heuristic interpretation of negative energy states in the dirac equation
this is not the end of the story . in 1929
oskar klein stumbled into an apparent paradox when trying to describe the scattering of a relativistic electron by a square potential using dirac s wave equation @xcite ( for pedagogical reviews see @xcite ) . in order to capture the essence of the problem without entering into unnecessary complication
we will study klein s paradox in the context of the klein - gordon equation .
= 5.0truein let us consider a square potential with height @xmath37 of the type showed in fig .
[ fig : potential ] . a solution to the wave equation in regions i and ii
is given by @xmath38 where the mass - shell condition implies that @xmath39 the constants @xmath40 and @xmath41 are computed by matching the two solutions across the boundary @xmath42 . the conditions @xmath43 and @xmath44 imply that @xmath45 at first sight one would expect a behavior similar to the one encountered in the nonrelativistic case .
if the kinetic energy is bigger than @xmath46 both a transmitted and reflected wave are expected , whereas when the kinetic energy is smaller than @xmath46 one only expect to find a reflected wave , the transmitted wave being exponentially damped within a distance of a compton wavelength inside the barrier .
indeed this is what happens if @xmath47 . in this case
both @xmath48 and @xmath49 are real and we have a partly reflected , and a partly transmitted wave . in the same way , if @xmath50 _ and _ @xmath51 then @xmath49 is imaginary and there is total reflection .
however , in the case when @xmath52 and the energy is in the range @xmath53 a completely different situation arises
. in this case one finds that both @xmath48 and @xmath49 are real and therefore the incoming wave function is partially reflected and partially transmitted across the barrier .
this is a shocking result , since it implies that there is a nonvanishing probability of finding the particle at any point across the barrier with negative kinetic energy ( @xmath54 ) !
this weird result is known as klein s paradox . as with the negative energy states , the klein paradox results from our insistence in giving a single - particle interpretation to the relativistic wave function
actually , a multiparticle analysis of the paradox @xcite shows that what happens when @xmath55 is that the reflection of the incoming particle by the barrier is accompanied by the creation of pairs particle - antiparticle out of the energy of the barrier ( notice that for this to happen it is required that @xmath52 , the threshold for the creation of a particle - antiparticle pair ) .
actually , this particle creation can be understood by noticing that the sudden potential step in fig .
[ fig : potential ] localizes the incoming particle with mass @xmath56 in distances smaller than its compton wavelength @xmath57 .
this can be seen by replacing the square potential by another one where the potential varies smoothly from @xmath58 to @xmath52 in distances scales larger than @xmath59 .
this case was worked out by sauter shortly after klein pointed out the paradox @xcite .
he considered a situation where the regions with @xmath60 and @xmath61 are connected by a region of length @xmath62 with a linear potential @xmath63 .
when @xmath64 he found that the transmission coefficient is exponentially small ) we will see how , in the case of the dirac field , this exponential behavior can be associated with the creation of electron - positron pairs due to a constant electric field ( schwinger effect ) . ] .
the creation of particles is impossible to avoid whenever one tries to locate a particle of mass @xmath56 within its compton wavelength .
indeed , from heisenberg uncertainty relation we find that if @xmath65 , the fluctuations in the momentum will be of order @xmath66 and fluctuations in the energy of order @xmath67 can be expected .
therefore , in a relativistic theory , the fluctuations of the energy are enough to allow the creation of particles out of the vacuum . in the case of a spin-@xmath68 particle ,
the dirac sea picture shows clearly how , when the energy fluctuations are of order @xmath56 , electrons from the dirac sea can be excited to positive energy states , thus creating electron - positron pairs . = 3.5truein it is possible to see how the multiparticle interpretation is forced upon us by relativistic invariance . in non - relativistic quantum mechanics observables
are represented by self - adjoint operator that in the heisenberg picture depend on time .
therefore measurements are localized in time but are global in space .
the situation is radically different in the relativistic case .
because no signal can propagate faster than the speed of light , measurements have to be localized both in time and space .
causality demands then that two measurements carried out in causally - disconnected regions of space - time can not interfere with each other . in mathematical terms
this means that if @xmath69 and @xmath70 are the observables associated with two measurements localized in two causally - disconnected regions @xmath71 , @xmath72 ( see fig .
[ fig : causal ] ) , they satisfy @xmath73=0 , \hspace*{1 cm } \mbox{if $ ( x_{1}-x_{2})^2<0 $ , for all $ x_1\in r_{1}$ , $ x_{2}\in r_{2}$}. \label{commutator}\end{aligned}\ ] ] hence , in a relativistic theory , the basic operators in the heisenberg picture must depend on the space - time position @xmath74 .
unlike the case in non - relativistic quantum mechanics , here the position @xmath75 _ is not _ an observable , but just a label , similarly to the case of time in ordinary quantum mechanics .
causality is then imposed microscopically by requiring @xmath76=0 , \hspace*{1 cm } \mbox{if $ ( x - y)^2<0$}. \label{microcausality}\end{aligned}\ ] ] a smeared operator @xmath77 over a space - time region @xmath40 can then be defined as @xmath78 where @xmath79 is the characteristic function associated with @xmath40 , @xmath80 eq .
( [ commutator ] ) follows now from the microcausality condition ( [ microcausality ] )
. therefore , relativistic invariance forces the introduction of quantum fields .
it is only when we insist in keeping a single - particle interpretation that we crash against causality violations . to illustrate the point ,
let us consider a single particle wave function @xmath33 that initially is localized in the position @xmath81 @xmath82 evolving this wave function using the hamiltonian @xmath83 we find that the wave function can be written as @xmath84 integrating over the angular variables , the wave function can be recast in the form @xmath85 the resulting integral can be evaluated using the complex integration contour @xmath86 shown in fig .
[ fig : contour ] .
the result is that , for any @xmath87 , one finds that @xmath88 for any @xmath75 .
if we insist in interpreting the wave function @xmath33 as the probability density of finding the particle at the location @xmath75 in the time @xmath89 we find that the probability leaks out of the light cone , thus violating causality .
we have learned how the consistency of quantum mechanics with special relativity forces us to abandon the single - particle interpretation of the wave function . instead we have to consider quantum fields whose elementary excitations are associated with particle states , as we will see below . in any scattering experiment ,
the only information available to us is the set of quantum number associated with the set of free particles in the initial and final states .
ignoring for the moment other quantum numbers like spin and flavor , one - particle states are labelled by the three - momentum @xmath18 and span the single - particle hilbert space @xmath90 @xmath91 the states @xmath92 form a basis of @xmath90 and therefore satisfy the closure relation @xmath93 the group of spatial rotations acts unitarily on the states @xmath94 .
this means that for every rotation @xmath95 there is a unitary operator @xmath96 such that @xmath97 where @xmath98 represents the action of the rotation on the vector @xmath99 , @xmath100 . using a spectral decomposition ,
the momentum operator @xmath101 can be written as @xmath102 with the help of eq .
( [ eq : transformation ] ) it is straightforward to check that the momentum operator transforms as a vector under rotations : @xmath103 where we have used that the integration measure is invariant under so@xmath104 .
since , as we argued above , we are forced to deal with multiparticle states , it is convenient to introduce creation - annihilation operators associated with a single - particle state of momentum @xmath18 @xmath105=\delta(\vec{p}-\vec{p}\ , ' ) , \hspace*{1 cm } [ a(\vec{p}),a(\vec{p}\,')]= [ a^{\dagger}(\vec{p}),a^{\dagger}(\vec{p}\,')]=0,\end{aligned}\ ] ] such that the state @xmath94 is created out of the fock space vacuum @xmath106 ( normalized such that @xmath107 ) by the action of a creation operator @xmath108 @xmath109 covariance under spatial rotations is all we need if we are interested in a nonrelativistic theory . however in a relativistic quantum field theory we must preserve more that so@xmath104 , actually we need the expressions to be covariant under the full poincar group iso@xmath110 consisting in spatial rotations , boosts and space - time translations . therefore , in order to build the fock space of the theory we need two key ingredients : first an invariant normalization for the states , since we want a normalized state in one reference frame to be normalized in any other inertial frame . and secondly a relativistic invariant integration measure in momentum space ,
so the spectral decomposition of operators is covariant under the full poincar group .
let us begin with the invariant measure .
given an invariant function @xmath111 of the four - momentum @xmath112 of a particle of mass @xmath56 with positive energy @xmath113 , there is an integration measure which is invariant under proper lorentz transformations are introduced for later convenience . ]
@xmath114 where @xmath115 represent the heaviside step function .
the integration over @xmath116 can be easily done using the @xmath117-function identity @xmath118=\sum_{x_i={\rm zeros\,\,of\,\,}f}{1\over |f'(x_{i})| } \delta(x - x_{i } ) , \label{eq : delta_identity}\end{aligned}\ ] ] which in our case implies that @xmath119 the second term in the previous expression correspond to states with negative energy and therefore does not contribute to the integral .
we can write then @xmath120 hence , the relativistic invariant measure is given by @xmath121 once we have an invariant measure the next step is to find an invariant normalization for the states .
we work with a basis @xmath122 of eigenstates of the four - momentum operator @xmath123 @xmath124 since the states @xmath125 are eigenstates of the three - momentum operator we can express them in terms of the non - relativistic states @xmath94 that we introduced in eq .
( [ eq : nonrelativistic_states ] ) @xmath126 with @xmath127 a normalization to be determined now . the states
@xmath128 form a complete basis , so they should satisfy the lorentz invariant closure relation @xmath129 at the same time , this closure relation can be expressed , using eq .
( [ eq : normalization ] ) , in terms of the nonrelativistic basis of states @xmath92 as @xmath130 ) for the nonrelativistic states , expression ( [ eq : closure_r ] ) follows provided @xmath131 ) so that @xmath132 is real , we define the lorentz invariant states @xmath125 as latexmath:[\ ] ] which proves that @xmath1353 $ ] and therefore @xmath1356 whenever the representation is real or pseudoreal .
since the gauge anomaly in four dimensions is proportional to @xmath1352 this means that anomalies appear only when the fermions transform in a complex representation of the gauge group .
n. dombey and a. calogeracos , _
seventy years of the klein paradox _
, phys . rept .
* 315 * ( 1999 ) 41 .
+ n. dombey and a. calogeracos , _ history and physics of the klein paradox _ ,
* 40 * ( 1999 ) 313 ( quant - ph/9905076 ) .
f. sauter , _ zum kleinschen paradoxon _
, z. phys .
* 73 * ( 1932 ) 547 .
h. b. g. casimir , _ on the attraction between two perfectly conducting plates _
akad . wet .
* 60 * ( 1948 ) 793 .
g. plunien , b. mller and w. greiner , _ the casimir effect _ ,
* 134 * ( 1986 ) 87 .
+ k. a. milton , _ the casimir effect : physical manifestation of zero - point energy _ , ( hep - th/9901011 ) .
+ k. a. milton , _ the casimir effect : recent controversies and progress
_ , j. phys . * a37 * ( 2004 ) r209 ( hep - th/0406024 ) .
+ s. k. lamoreaux , _ the casimir force : background , experiments , and applications _ , rep .
* 68 * ( 2005 ) 201 .
m. j. sparnaay , _ measurement of attractive forces between flat plates _ , physica * 24 * ( 1958 ) 751 . m. abramowitz and i. a. stegun , _ handbook of mathematical functions _ ,
dover 1972 .
y. aharonov and d. bohm , _ significance of the electromagnetic potentials in the quantum theory _
, phys . rev .
* 115 * ( 1955 ) 485 .
a. m. dirac , _ quantised singularities in the electromagnetic field _ , proc .
* 133 * ( 1931 ) 60 .
s. adler , _ axial - vector vertex in spinor electrodynamics , _ phys .
* 177 * ( 1969 ) 2426 .
+ j. s. bell and r. jackiw , _ a pcac puzzle : @xmath951 in the sigma model _ , nuovo cimento * a60 * ( 1969 ) 47 . j. steinberger , _ on the use of substraction fiels and the lifetimes of some types of meson decay _
, phys . rev .
* 76 * ( 1949 ) 1180 .
f. j. yndurin , _ the theory of quark and gluon interactions _ , springer 1999 .
g. t hooft , _ how the instantons solve the u(1 ) problem _ , phys . rept .
* 142 * ( 1986 ) 357 .
d. g. sutherland , _ current algebra and some nonstrong mesonic decays _
, nucl . phys . *
b2 * ( 1967 ) 433 .
+ m. j. g. veltman , _ theoretical aspects of high - energy neutrino interactions _ , proc .
* a301 * ( 1967 ) 107 .
s. l. adler and w. a. bardeen , _ absence of higher order corrections in the anomalous axial vector divergence equation _
* 182 * ( 1969 ) 1517 .
e. witten , _
an su(2 ) anomaly _ , phys .
* b117 * ( 1982 ) 324 .
s. eidelman et al .
_ review of particle phhysics _
, phys . lett .
* b592 * ( 2004 ) 1 ( http://pdg.lbl.gov ) .
d. j. gross and f. wilczek , _ ultraviolet behavior of nonabelian gauge theories _ ,
lett . * 30 * ( 1973 ) 1343
. h. d. politzer , _ reliable perturbative results for strong interations ? _ , phys .
30 * ( 1973 ) 1346 . g. t hooft , remarks at the _ colloquium on renormalization of yang - mills fields and applications to particle physics _ , marseille 1972 . i. b. khriplovich , _
green s functions in theories with a non - abelian gauge group _ , yad . fiz .
* 10 * ( 1969 ) 409 [ sov
. phys . * 10 * ( 1970 ) 235 ] .
+ m. v. terentiev and v. s. vanyashin , _ the vacuum polarization of a charged vector field _ , zh . eskp .
* 48 * ( 1965 ) 565 [ sov
jetp * 21 * ( 1965 ) 375 ] .
k. g. wilson , _ renormalization group and critical phenomena 1
. renormalization group and the kadanoff scaling picture , _ phys . rev . * b4 * ( 1971 ) 3174 .
+ k. g. wilson , _ renormalization group and critical phenomena 2
. phase space cell analysis of critical behavior _ , phys . rev . * b4 * ( 1971 ) 3184 + k. g. wilson , _ the renormalization group and critical phenomena _ , rev .
* 55 * ( 1983 ) 583 .
l. p. kadanoff , _ scaling laws for ising models near @xmath1359 _ , physics * 2 * ( 1966 ) 263 .
j. schwinger , _ on gauge invariance and vacuum polarization _ ,
phys . rev .
* 82 * ( 1951 ) 664 .
e. brezin and c. itzykson , _ pair production in vacuum by an alternating field _
, phys . rev . *
d2 * ( 1970 ) 1191 .
s. w. hawking , _ particle creation by black holes _
* 43 * ( 1975 ) 199 .
m. k. parikh and f. wilczek , _ hawking radiation as tunneling _ ,
* 85 * ( 2000 ) 5042 ( hep - th/9907001 ) yu . a. golfand and e. p. likhtman , _ extension of the algebra of poincar group generators and violations of p - invariance _ , jetp lett .
* 13 * ( 1971 ) 323 .
+ d. v. volkov and v. p. akulov , _ is the neutrino a goldstone particle _ , phys .
b46 * ( 1973 ) 109 .
+ j. wess and b. zumino , _ a lagrangian model invariant under supergauge transformations _ ,
b49 * ( 1974 ) 52 .
r. haag , j. opuszaski and m. sohnius , _ all possible generators of supersymmetries of the s - matrix _ , nucl . phys . *
b88 * ( 1975 ) 257 . | in these lectures we present a few topics in quantum field theory in detail .
some of them are conceptual and some more practical .
they have been selected because they appear frequently in current applications to particle physics and string theory . |
in @xcite , bergman defined the _ logarithmic limit - set _ of an algebraic variety in order to study its exponential behavior at infinity .
we follow @xcite in calling this set the _ bergman complex _ of the variety .
bergman complexes have recently received considerable attention in several areas , such as tropical algebraic geometry and dynamical systems .
they are the _ non - archimedean amoebas _ of @xcite and the _ tropical varieties _ of @xcite .
when the variety is a linear space , so that the defining ideal @xmath4 is generated by linear forms , sturmfels @xcite showed that the bergman complex can be described solely in terms of the matroid associated to the linear ideal .
he used this description to define the bergman complex @xmath10 of an arbitrary matroid @xmath1 .
ardila and klivans @xcite showed that , appropriately subdivided , the bergman complex of a matroid @xmath1 is the order complex of the proper part of the lattice of flats @xmath11 of the matroid .
this result implies that the bergman complex of an arbitrary matroid @xmath1 is a finite , pure polyhedral complex , which is homotopy equivalent to a wedge of spheres .
total positivity is another topic which has received a great deal of recent interest .
although the classical theory concerns matrices in which all minors are positive , in the past decade this theory has been extended by lusztig @xcite , who introduced the totally positive variety @xmath12 in an arbitrary reductive group @xmath13 and the totally positive part @xmath14 of a real flag variety @xmath15 .
more recently , the positive part of the tropicalization of an affine variety ( or positive tropical variety , for short ) was introduced by speyer and williams @xcite .
sturmfels @xcite suggested the notion of a positive bergman complex @xmath0 of an oriented matroid @xmath1 and conjectured its relation to the las vergnas face lattice of @xmath1 .
we define the positive bergman complex and positive bergman fan so that given a linear ideal @xmath4 with associated oriented matroid @xmath5 , the positive tropical variety associated to @xmath4 is equal to the positive bergman fan of @xmath5 .
we give a proof of sturmfels conjecture : appropriately subdivided , @xmath0 is a geometric realization of the order complex of the proper part of the las vergnas face lattice of @xmath1 .
@xmath0 is homeomorphic to a sphere and naturally sits inside @xmath2 , the bergman complex of the underlying unoriented matroid of @xmath1 .
we conclude by showing that , for the oriented matroid of the complete graph @xmath7 , the face poset of a certain coarse " subdivision of @xmath8 is dual to the face poset of the associahedron @xmath9 .
the paper is organized as follows . in section [ sectionmw ]
we introduce a certain oriented matroid @xmath16 which will play an important role in our work . in section [ posberg ]
we define the positive bergman complex and prove our main theorem . in section [ postrop ]
we explain the relation between the positive bergman complex of an oriented matroid and the positive tropical variety of a linear ideal . in section [ top ]
we describe the topology of the positive bergman complex of an oriented matroid .
finally , in sections [ kn ] and [ hooklength ] we describe in detail the positive bergman complex of the oriented matroid of @xmath7 : we relate it to the associahedron , and we give a formula for the number of full - dimensional fine cells within a full - dimensional coarse cell . throughout this paper
we will abuse notation and use @xmath1 to denote either a matroid or oriented matroid , depending on the context .
similarly , we will use the term circuits " to describe either unsigned or signed circuits .
when the distinction between matroids and oriented matroids is important , we will use @xmath3 to denote the underlying matroid of an oriented matroid @xmath1 .
let @xmath1 be an oriented matroid on the ground set @xmath18=\{1 , 2 , \dots , n\}$ ] whose collection of signed circuits is @xmath19 .
let @xmath20 and regard @xmath21 as a weight function on @xmath18 $ ] .
for any circuit @xmath22 define @xmath23 to be the @xmath21-maximal subset of the circuit @xmath24 in other words , the collection of elements of @xmath24 which have the largest weight .
we will say that the circuit @xmath24 _ achieves its largest weight with respect to @xmath21 _ at @xmath23 . define @xmath25 to be the collection of inclusion - minimal sets of the collection @xmath26 .
we then define @xmath16 to be the oriented matroid on @xmath18 $ ] whose collection of circuits is @xmath25 .
it is not clear that @xmath16 is a well - defined oriented matroid ; we will prove this shortly .
given @xmath20 , let @xmath27 denote the unique flag of subsets @xmath28 $ ] such that @xmath21 is constant on each set @xmath29 and satisfies @xmath30 .
we call @xmath27 the _ flag _ of @xmath21 , and we say that the _ weight class _ of @xmath21 or of the flag @xmath31 is the set of vectors @xmath32 such that @xmath33 .
it is clear that @xmath16 depends only on the flag @xmath34 and so we also refer to this oriented matroid as @xmath35 .
[ mainexample ] let @xmath1 be the oriented matroid of the digraph @xmath36 shown in figure [ d ] .
equivalently , let @xmath1 be the oriented matroid of the point configuration shown in figure [ point ] . , height=96 ]
note that @xmath36 is an acyclic orientation of @xmath37 , the complete graph on @xmath38 vertices .
the signed circuits @xmath19 of @xmath1 are @xmath39 , together with the negatives of every set of this collection .
choose @xmath21 such that @xmath40 , which corresponds to the flag @xmath41 .
if we calculate @xmath23 for each @xmath22 we get @xmath42 , together with the negatives of every set of this collection .
however , @xmath43 is not inclusion - minimal in this collection , as it contains @xmath44 and @xmath45 .
thus @xmath25 is equal to @xmath46 , together with the negatives of every set , and @xmath16 is the oriented matroid whose collection of signed circuits is @xmath25 .
notice that in this case @xmath16 is the oriented matroid of the digraph @xmath47 in figure [ mw ] .
.,height=96 ] we must show that @xmath16 is well - defined . for convenience ,
we review here the _ circuit axioms for oriented matroids _
@xcite : 1 .
@xmath48 is not a signed circuit .
[ c1 ] 2 . if @xmath49 is a signed circuit , then so is @xmath50 .
[ c2 ] 3 .
no proper subset of a circuit is a circuit .
[ c3 ] 4 .
if @xmath51 and @xmath52 are circuits with @xmath53 and @xmath54 , then there is a third circuit @xmath55 with @xmath56 and @xmath57 .
[ c4 ] we will also need the following stronger characterization of oriented matroids : ( * ? ? ?
* theorem 3.2.5 ) let @xmath19 be a collection of signed subsets of a set @xmath58 satisfying c1 , c2 , c3 .
then c4 is equivalent to c4 : 1 . for all @xmath59 , @xmath60 and @xmath61 ,
there is a @xmath62 such that @xmath63 , @xmath64 , and @xmath65 .
[ c4 ] let @xmath1 be an oriented matroid on @xmath18 $ ] and @xmath20
. then @xmath16 is an oriented matroid .
the strategy of our proof is to show that if c1 , c2 , c3 and c4 hold for @xmath1 , then c1 , c2 , c3 and c4 hold for @xmath16 .
first note that it is obvious that c1 and c2 hold for @xmath16 .
c3 holds for @xmath16 because we defined @xmath25 to consist of _ inclusion - minimal _ elements .
it remains to show that c4 holds for @xmath16 . to do this , we start with two circuits in @xmath16 , lift them to circuits in @xmath1 , and then use c4 for @xmath1 to show that c4 holds for @xmath16 . take @xmath66 and @xmath67 in @xmath25 such that @xmath68 and @xmath69 . by definition ,
there exist circuits @xmath51 and @xmath52 of @xmath1 such that @xmath70 and @xmath71 .
notice that the presence of @xmath72 in @xmath66 and @xmath67 guarantees that the maximum weights occurring in @xmath51 and in @xmath52 are both equal to @xmath73 .
choose any @xmath74 . clearly such an @xmath75 exists . then @xmath76 .
by c4 for @xmath1 , there exists a circuit @xmath49 in @xmath19 such that 1 .
@xmath77 , 2 .
@xmath78 , and 3 .
@xmath79 .
look at @xmath80 .
we will prove that @xmath80 contains the third circuit of @xmath81 which we are looking for .
we want to show that 1 .
@xmath82 2 .
@xmath83 first , it is obvious that @xmath72 is not in @xmath84 , since @xmath72 was not in @xmath85 . clearly @xmath84 is a subset of @xmath86 . to show @xmath87 and @xmath88
, we just need to show that the maximum weight which occurs in @xmath49 is also equal to @xmath73 . by @xmath89 and @xmath90 ,
this maximum weight is at most @xmath73 .
by @xmath91 , equality is attained for @xmath92 , since @xmath93 .
note that if @xmath80 is not inclusion - minimal in the set @xmath94 , then it contains some inclusion - minimal @xmath95 for another circuit @xmath96 of @xmath1 . and
since @xmath82 and @xmath83 , it is clear that we also have @xmath97 and @xmath98 . in the following proposition , we describe the bases of @xmath16 and their orientations .
let us say that if @xmath99 $ ] and @xmath100 , the _
@xmath21-weight _ of @xmath101 is the sum @xmath102 .
the bases of @xmath16 are the bases of @xmath1 which have minimal @xmath21-weight .
the basis orientations of @xmath81 are equal to their orientations in @xmath1 .
we know that if @xmath103 is an oriented matroid on @xmath18 $ ] with signed circuits @xmath19 , then the bases of @xmath103 are the maximal subsets of @xmath18 $ ] which contain no circuit .
thus , the bases @xmath104 of @xmath16 are the maximal subsets of @xmath18 $ ] which do not contain a set of the collection @xmath105 .
we want to show that @xmath104 is exactly the set of bases of @xmath1 which have minimal @xmath21-weight .
first let us choose a basis @xmath15 of @xmath1 which has minimal @xmath21-weight .
we claim that @xmath15 is independent in @xmath16 .
suppose that @xmath15 contains a subset of the form @xmath23 , where @xmath24 is a circuit of @xmath1 .
write @xmath106 , and choose any @xmath107 .
we now construct a new basis @xmath108 of @xmath1 of smaller weight than @xmath15 , as follows .
start with the set @xmath24 .
we know that @xmath109 is a spanning set of @xmath1 .
let @xmath110 be a minimal subset ( possibly empty ) of @xmath111 such that @xmath112 is still a spanning set of @xmath1 .
since @xmath24 is minimally dependent in @xmath1 , the set @xmath113 will be a basis of @xmath1 .
now by the basis exchange axiom , for some @xmath114 , @xmath115 is a basis of @xmath1 .
since @xmath116 , it follows that @xmath117 .
thus , the weight of @xmath118 is strictly less than the weight of @xmath119 , which implies that the weight of @xmath108 is less than the weight of @xmath15 .
this is a contradiction .
the previous argument shows that @xmath120 .
but every circuit of @xmath81 is contained in a circuit of @xmath1 , so @xmath121 .
it follows that @xmath1 and @xmath81 have the same rank .
now let @xmath15 be a basis of @xmath81 , _
i.e. _ @xmath15 is a maximal subset of @xmath18 $ ] which does not contain an element of @xmath122 .
we claim that @xmath15 is a basis of @xmath1 with minimal @xmath21-weight .
first note that @xmath15 is clearly independent in @xmath1 : if it were dependent in @xmath1 , it would contain some circuit @xmath24 of @xmath1 and hence would contain @xmath23 . also , @xmath15 has @xmath123 elements . therefore it is a basis of @xmath1 . finally , let us show that @xmath15 has minimal @xmath21-weight .
suppose not .
let @xmath124 be the elements of @xmath15 with highest weight .
we claim that @xmath125 for some circuit @xmath24 of @xmath1 , which will be a contradiction .
since @xmath15 is a basis of @xmath1 , adding to @xmath15 any element of @xmath18\setminus b$ ] creates a circuit .
since @xmath15 is not a basis of minimal @xmath21-weight , there must be an element @xmath126\setminus b$ ] such that the weight of @xmath118 is strictly less than the weight of each of the elements @xmath124 .
thus @xmath127 contains a circuit @xmath24 , and @xmath128 , as claimed . to prove the claim about orientations , start with a basis @xmath15 of minimal weight of @xmath1 .
recall that an oriented matroid has exactly two basis orientations , which are opposite to each other .
therefore we can assume without loss of generality that @xmath15 has the same orientation in @xmath1 and @xmath81 .
for any two ordered bases @xmath129 and @xmath130 of @xmath81 with @xmath131 , we have @xmath132 where @xmath133 is the chirotope of @xmath81 , and @xmath134 is one of the two opposite signed circuits of @xmath81 in @xmath135 . now @xmath136 and
@xmath137 are also bases of @xmath1 ; let @xmath24 be one of the two opposite signed circuits of @xmath1 in @xmath138 . then @xmath139 contains a circuit of @xmath81 ; it must be either @xmath134 or @xmath140 . in any case
, we have @xmath141 , so @xmath142 .
it follows that if @xmath136 has the same orientation in @xmath1 and @xmath81 , then so does @xmath137 .
recall that one can obtain any basis of a matroid from any other by a sequence of simple basis exchanges of the type above .
since @xmath15 has the same orientation in @xmath1 and @xmath81 , so does any other basis of @xmath81 .
our goal in this section is to define the positive bergman complex of an oriented matroid @xmath1 and to relate it to the las vergnas face lattice of @xmath1 , thus answering sturmfels question @xcite .
we begin by giving some background on the bergman complex and fan of a ( unoriented ) matroid .
the _ bergman fan _ of a matroid @xmath1 on the ground set @xmath18 $ ] is the set @xmath143 the _ bergman complex _ of @xmath1 is @xmath144 where @xmath145 is the sphere @xmath146 for simplicity , in this section we will concentrate on the bergman complex of @xmath1 , but similar arguments hold for the bergman fan of @xmath1 .
since the matroid @xmath81 depends only on the weight class that @xmath21 is in , the bergman complex of @xmath1 is a disjoint union of the weight classes of flags @xmath31 such that @xmath147 has no loops .
we say that the weight class of a flag @xmath31 is _ valid _ for @xmath1 if @xmath147 has no loops .
there are two polyhedral subdivisions of @xmath10 , one of which is clearly finer than the other . the _ fine subdivision _ of @xmath10 is the subdivision of @xmath10 into valid weight classes : two vectors @xmath148 and @xmath149 of @xmath10 are in the same class if and only if @xmath150 . the _ coarse subdivision _ of @xmath10 is the subdivision of @xmath10 into @xmath81-equivalence classes : two vectors @xmath148 and @xmath149 of @xmath10 are in the same class if and only if @xmath151 .
the following results give alternative descriptions of @xmath10 : @xcite[cf ] given an ( unoriented ) matroid @xmath1 on the ground set @xmath18 $ ] and @xmath20 which corresponds to a flag @xmath34 , the following are equivalent : 1 .
@xmath35 has no loops . 2 . for each circuit @xmath24 of @xmath1
, @xmath139 contains at least two elements of @xmath24 .
3 . @xmath31 is a flag of flats of @xmath1 .
@xcite[cf2 ] let @xmath1 be a ( unoriented ) matroid .
then the fine subdivision of the bergman complex @xmath10 is a geometric realization of @xmath152 , the order complex of the proper part of the lattice of flats of @xmath1 .
we are now ready for the positive analogues of these concepts .
the _ positive bergman fan _ of an oriented matroid @xmath1 on the ground set @xmath18 $ ] is @xmath153 the _ positive bergman complex _ of @xmath1 is @xmath154 within each equivalence class of the coarse subdivision of @xmath10 , the vectors @xmath21 give rise to the same unoriented @xmath81 .
since the orientation of @xmath81 is inherited from that of @xmath1 , they also give rise to the same oriented matroid @xmath81 .
therefore each coarse cell of @xmath155 is either completely contained in or disjoint from @xmath0 .
thus @xmath0 inherits the coarse and the fine subdivisions from @xmath10 , and each subdivision of @xmath0 is a subcomplex of the corresponding subdivision of @xmath10 .
let @xmath1 be an acyclic oriented matroid on the ground set @xmath18 $ ] .
we say that a covector @xmath156 of @xmath1 is _ positive _ if each of its entries is @xmath157 or @xmath158 .
we say that a flat of @xmath1 is _ positive _ if it is the @xmath158-set of a positive covector .
additionally , we consider the set @xmath18 $ ] to be a positive flat .
for example , if @xmath1 is the matroid of example [ mainexample ] , then @xmath159 is a positive flat which is the @xmath158-set of the positive covector @xmath160 .
las vergnas face lattice _
@xmath161 is the lattice of positive flats of @xmath1 , ordered by containment .
note that the lattice of positive flats of the oriented matroid @xmath1 sits inside @xmath11 , the lattice of flats of @xmath1 .
let @xmath1 be the oriented matroid from example [ mainexample ] .
the positive covectors of @xmath1 are @xmath162 and the positive flats are @xmath163 . the lattice of positive flats of @xmath1 is shown in figure [ posflats ] , alongside the lattice of flats of @xmath1 .
we now give an analogue of theorem [ cf ] .
[ maintheorem ] given an oriented matroid @xmath1 and @xmath20 which corresponds to a flag @xmath34 , the following are equivalent : 1 .
@xmath35 is acyclic .
[ 1 ] 2 . for each signed circuit @xmath24 of @xmath1 , @xmath139 contains a positive element and a negative element of @xmath24 . [ 2 ] 3 .
@xmath31 is a flag of positive flats of @xmath1 .
[ 3 ] first we will show that [ 1 ] and [ 2 ] are equivalent .
the statement that @xmath16 is acyclic means that @xmath16 has no all - positive circuit : in other words , each circuit of @xmath16 contains a positive and a negative term .
since @xmath16 is the matroid whose circuits are the inclusion - minimal elements of the set @xmath94 , this means that for each circuit @xmath24 of @xmath1 , @xmath164 contains a positive and a negative term .
finally , this is equivalent to the statement that for each circuit @xmath24 of @xmath1 , @xmath24 achieves its maximum value with respect to @xmath21 on both @xmath165 and @xmath166 .
next we show that [ 3 ] implies [ 2 ] .
assume we have an @xmath21 such that @xmath31 is a flag of positive flats .
let the flats of this flag be @xmath167 .
for each @xmath168 , @xmath169 - f_i)$ ] is a positive covector . by orthogonality of circuits and covectors , we know that for any circuit @xmath24 and any covector @xmath170 , @xmath171 and @xmath172 are either both empty or both non - empty . for any circuit @xmath24 of @xmath1 , consider the largest @xmath173 such that @xmath174 - f_i)$ ] is non - empty .
then clearly @xmath24 will attain its maximum on @xmath175 and @xmath23 contains a positive element and a negative element of @xmath24 .
finally , assume that [ 1 ] and [ 2 ] hold , but [ 3 ] does not .
from [ 1 ] we know that @xmath147 is acyclic ; therefore the unoriented @xmath147 has no loops , and @xmath31 is a flag of flats by theorem [ cf ] .
let @xmath168 be a flat which is not positive ; by ( * ? ? ?
* proposition 9.1.2 ) this is equivalent to saying that @xmath176 is not acyclic .
let @xmath24 be a positive circuit of @xmath176 ; then we can find a circuit @xmath49 of @xmath1 such that @xmath177 .
then @xmath49 has positive elements of weight greater than @xmath178 , and no negative elements of weight greater than @xmath178 .
it follows that @xmath179 is positive , contradicting [ 2 ] .
[ lasvergnas ] let @xmath1 be an oriented matroid .
then the fine subdivision of @xmath0 is a geometric realization of @xmath180 , the order complex of the proper part of the las vergnas face lattice of @xmath1 .
in @xcite , the notion of the positive part of the tropicalization of an affine variety ( or positive tropical variety , for short ) was introduced , an object which has the structure of a polyhedral fan in @xmath181 . in order to describe this object
, we must define an _
initial ideal_. let @xmath182 $ ] and @xmath20 . if @xmath183 , define the _ initial form _
@xmath184 to be the sum of all terms @xmath185 such that the inner product @xmath186 is maximal . for an ideal @xmath4 of @xmath187 $ ] ,
define the _ initial ideal _
@xmath188 to be the ideal generated by @xmath189 for all @xmath190 .
if @xmath4 is an ideal in a polynomial ring with @xmath191 variables , the positive tropical variety associated to @xmath4 is denoted by @xmath192 and can be characterized as follows : @xmath193$}\}.\ ] ] now recall that if @xmath4 is a linear ideal ( an ideal generated by linear forms ) , we can associate to it an oriented matroid @xmath5 as follows .
write each linear form @xmath194 in the form @xmath195 , where @xmath196 for all @xmath173 .
we then define @xmath5 to be the oriented matroid whose set of signed circuits consists of all minimal collections of the form @xmath197 .
we now prove the following easy statement .
if @xmath4 is a linear ideal and @xmath5 is the associated oriented matroid , then @xmath198 .
@xmath199\ } \\ & = & \operatorname{trop}^+ v(i).\end{aligned}\ ] ]
the topology of the positive bergman complex of an oriented matroid is very simple : it is homeomorphic to a sphere .
this follows from corollary [ lasvergnas ] together with results about the las vergnas face lattice , which we will review here .
* theorem 4.3.5 ) [ lv ] let @xmath1 be an acyclic oriented matroid of rank @xmath200
. then the las vergnas lattice @xmath161 is isomorphic to the face lattice of a @xmath201 regular cell decomposition of the @xmath202-sphere .
* proposition 4.7.8 ) [ cell ] let @xmath203 be a regular cell complex
. then its geometric realization is homeomorphic to the geometric realization of the order complex of its face poset .
the previous two results imply that the geometric realization of the order complex of the las vergnas lattice is homeomorphic to a sphere . putting this together with corollary [ lasvergnas ]
, we get the following result .
the positive bergman complex of an oriented matroid is homeomorphic to a sphere .
in this section , we wish to describe the positive bergman complex @xmath8 of the graphical oriented matroid @xmath204 of an acyclic orientation of the complete graph @xmath7 .
we start by reviewing the description of the bergman complex @xmath205 of the unoriented matroid @xmath204 , obtained in @xcite . for the moment
we need to consider @xmath7 as an unoriented graph .
an _ equidistant @xmath191-tree _ @xmath206 is a rooted tree with @xmath191 leaves labeled @xmath207 , and lengths assigned to each edge in such a way that the total distances from the root to each leaf are all equal .
the internal edges are required to have positive lengths .
figure [ figure : tree ] shows an example of an equidistant @xmath38-tree . to each equidistant @xmath191-tree @xmath206 we assign a distance vector @xmath208 : the distance @xmath209 is equal to the length of the path joining leaves @xmath173 and @xmath210 in @xmath206 .
figure [ figure : tree ] also shows the distance vector of the tree , regarded as a weight function on the edges of @xmath37 .
the bergman fan @xmath211 can be regarded as a space of equidistant @xmath191-trees , as the following theorem shows .
@xcite the distance vector of an equidistant @xmath191-tree , when regarded as a weight function on the edges of @xmath7 , is in the bergman fan @xmath211 .
conversely , any point in @xmath211 is the distance vector of a unique equidistant @xmath191-tree .
as mentioned earlier , the fine subdivision of @xmath212 is well understood for any matroid @xmath1 .
the following theorem shows that the coarse subdivision of @xmath211 also has a nice description : it is a geometric realization of the well - studied simplicial complex of trees @xmath213 , sometimes called the whitehouse complex @xcite .
@xcite let @xmath214 .
let @xmath206 and @xmath215 be the corresponding equidistant @xmath191-trees . the following are equivalent : 1 .
@xmath21 and @xmath216 are in the same cell of the coarse subdivision .
2 . @xmath206 and @xmath215 have the same combinatorial type .
now we return to the setting of oriented matroids .
the positive bergman complex @xmath8 is defined in terms of an acyclic orientation of @xmath7 .
this graph has @xmath217 acyclic orientations , corresponding to the @xmath217 permutations of @xmath18 $ ] .
the orientation corresponding to the permutation @xmath218 is given by @xmath219 for @xmath220 .
clearly the @xmath217 orientations of @xmath7 will give rise to positive bergman complexes which are equal up to relabeling .
therefore , throughout this section , the edges of @xmath7 will be oriented @xmath221 for @xmath220 . as we go around a cycle @xmath24 of @xmath7 , @xmath165 is the set of edges which are crossed in the forward direction , and @xmath166 is the set of edges which are crossed in the backward direction .
let @xmath21 be a weight vector on the edges of the oriented complete graph @xmath7 .
let @xmath206 be the corresponding equidistant tree .
the following are equivalent : 1 .
@xmath21 is in @xmath8 .
@xmath206 can be drawn in the plane without crossings in such a way that its leaves are numbered @xmath222 from left to right .
we add three intermediate steps to the equivalence : 1 .
_ in any cycle @xmath24 , the @xmath21-maximum is achieved in @xmath165 and @xmath166 . _ 2 .
_ in any triangle @xmath24 , the @xmath21-maximum is achieved in @xmath165 and @xmath166 . _
_ for any three leaves @xmath223 in @xmath206 , the leaf @xmath210 does not branch off before @xmath173 and @xmath224 ; _ i.e._,their branching order is one of the following : _ + [
figure : branching ] the equivalence @xmath225 follows from theorem [ maintheorem ] , and the implication @xmath226 is trivial .
now we show that @xmath227 .
proceed by contradiction .
consider a cycle @xmath228 , with @xmath224 minimal , such that @xmath229 is not satisfied .
consider the cycles @xmath230 and @xmath231 , which do satisfy @xmath229 .
since @xmath24 does not satisfy @xmath229 , the edge @xmath232 must be @xmath21-maximum in @xmath206 , along with another edge @xmath72 of the opposite orientation .
similarly , the edge @xmath233 must be @xmath21-maximum in @xmath234 , along with another edge @xmath75 of the opposite orientation . therefore , in @xmath24 , the edges @xmath72 and @xmath75 are @xmath21-maximum and have opposite orientations .
this is a contradiction .
let us now show @xmath235 . in triangle @xmath236
( where we can assume @xmath223 ) , @xmath237 holds if and only if we have one of the following : @xmath238 these three conditions correspond , in that order , to the three possible branching orders of @xmath239 and @xmath224 in @xmath206 prescribed by condition @xmath240 .
finally we show @xmath241 .
the backward implication is immediate .
we prove the forward implication by induction on @xmath191 .
the case @xmath242 is clear .
now let @xmath243 , and assume that condition @xmath240 holds .
consider a lowest internal node @xmath149 ; it is incident to several leaves , which must have consecutive labels @xmath244 by @xmath240 .
let @xmath215 be the tree obtained from @xmath206 by removing leaf @xmath173 .
this smaller tree satisfies @xmath240 , so it can be drawn in the plane with the leaves in order from left to right .
now we simply find node @xmath149 in this drawing , and attach leaf @xmath173 to it , putting it to the left of all the other leaves incident to @xmath149 .
this is a drawing of @xmath206 satisfying 2 .
the _ associahedron _
@xmath9 is a well - known @xmath245-dimensional polytope whose vertices correspond to planar rooted trees @xcite .
there is a close relationship between @xmath8 and @xmath9 .
the face poset of the coarse subdivision of @xmath8 , with a @xmath246 attached , is dual to the face poset of the associahedron @xmath9 . in the trees corresponding to the cells of @xmath8 , the labeling of the leaves always increases from left to right .
we can forget these labels and obtain the usual presentation of the dual to the associahedron , whose facets correspond to planar rooted trees .
figure [ assoc ] shows the positive bergman complex of @xmath37 ( in bold ) within the bergman complex of @xmath37 .
vertices of the coarse subdivision are shown as black circles ; vertices of the fine subdivision but not the coarse subdivision are shown as transparent circles . observe that the coarse subdivision of @xmath247 is a pentagon , whose face poset is the face poset of the associahedron @xmath248 ( which is self - dual ) . , height=192 ] now , recall that different orientations of @xmath7 give rise to different positive bergman complexes .
let us make two comments about the way in which these positive bergman complexes fit together .
consider the @xmath217 different acyclic orientations @xmath249 of @xmath7 , each corresponding to a permutation @xmath218 of @xmath18 $ ] .
each orientation @xmath249 gives rise to a positive bergman complex : it consists of those weight vectors such that the corresponding tree can be drawn with the leaves labeled @xmath250 from left to right .
clearly , each permutation and its reverse give the same positive bergman complex .
the @xmath251 possible positive bergman complexes @xmath8 give a covering of @xmath205 , and each one of them is dual to the associahedron @xmath9 .
this corresponds precisely to the known covering of the space of trees with @xmath251 polytopes dual to the associahedron , as described in @xcite .
also , recall from @xcite that the bergman complex @xmath205 is topologically a wedge of @xmath252 spheres .
it is now easy to realize this statement combinatorially as follows .
any tree can be drawn in the plane so that the leftmost leaf is labeled @xmath253 .
therefore , @xmath205 is covered by the positive bergman complexes corresponding to the @xmath252 permutations @xmath218 such that @xmath254 .
it follows that @xmath205 is the wedge of @xmath252 spherical polytopes dual to the associahedron @xmath9 .
since @xmath8 and @xmath205 are @xmath245-dimensional , we will call the @xmath245-dimensional cells inside them _ full - dimensional_. in this section we will give a formula reminiscent of the `` hook - length '' formula for the number of full - dimensional fine cells within a full - dimensional coarse cell of @xmath205 .
[ finecells ] let @xmath255 be a rooted binary tree with @xmath191 labeled leaves .
for each internal vertex @xmath149 of @xmath255 , let @xmath256 be the number of internal vertices of @xmath255 which are descendants of @xmath149 , including @xmath149 .
let @xmath257 be the coarse cell of @xmath205 corresponding to tree @xmath255 .
there are exactly @xmath258 full - dimensional fine cells in @xmath257 .
the cell @xmath257 consists of the distance vectors @xmath259 of all equidistant @xmath191-trees @xmath206 of combinatorial type @xmath255 .
notice that @xmath260 , where @xmath149 is the lowest common ancestor of leaves @xmath173 and @xmath210 in @xmath206 , @xmath261 is the distance from @xmath149 to the root of @xmath206 , and @xmath262 is the distance from the root of @xmath206 to any of its leaves . to specify a full - dimensional fine cell in @xmath257 , one needs to specify the relative order of the @xmath209 s .
equivalently , in the tree @xmath206 that @xmath263 comes from , one needs to specify the relative order of the heights of the internal vertices , consistently with the combinatorial type of tree @xmath255 .
therefore , the fine cells in @xmath257 correspond to the labellings of the @xmath264 internal vertices of @xmath255 with the numbers @xmath265 , such that the label of each vertex is smaller than the labels of its offspring . in the language of ( * ? ? ?
1.3 ) , these are precisely the _ increasing binary trees _ of type @xmath266 , where @xmath266 is the result of removing the leaves of tree @xmath255 , and the edges incident to them .
figure [ figure : increasing ] shows a tree type @xmath255 and one of the increasing binary trees of type @xmath266 .
let @xmath270 and @xmath271 be the left and right subtrees of @xmath266 .
let @xmath136 and @xmath137 be the events that @xmath270 and @xmath271 are labeled increasingly in @xmath268 , and let @xmath15 be the event that the root of @xmath255 is labeled @xmath253 . then @xmath272 .
it is clear that @xmath273 and @xmath137 are independent events .
it is also clear that @xmath274 and @xmath275 .
therefore , @xmath276 the result follows by induction .
we recall the known bijection between increasing binary trees with vertices labeled @xmath278 , and permutations of @xmath279 ( * ? ? ?
it is defined recursively : the permutation @xmath280 corresponding to the increasing binary tree @xmath206 is @xmath281 , where @xmath282 and @xmath283 are the left and right subtrees of @xmath206 .
for example , the tree of figure [ figure : increasing ] corresponds to the permutation @xmath284 .
it is not difficult to see how @xmath206 can be recovered uniquely from @xmath280 . to show the second result ,
recall that the bergman complex @xmath205 is covered by @xmath217 positive bergman complexes .
each permutation @xmath218 of @xmath18 $ ] gives rise to a positive bergman complex @xmath8 ; this complex parameterizes those trees which can be drawn in the plane so that its leaves are in the order prescribed by @xmath218 . with @xmath252 fine cells in each positive bergman complex
, we get a covering of @xmath205 with @xmath286 fine cells .
each fine cell appears several times in this covering , since it sits inside several positive bergman complexes .
more precisely , each binary tree with @xmath191 labeled leaves can be drawn in the plane in exactly @xmath287 ways : at each internal vertex , we may or may not switch the left and right subtrees . therefore , each fine cell of the bergman complex @xmath205 is inside @xmath287 different positive bergman complexes .
the desired result follows .
recall that the maximum - dimensional fine cells of @xmath205 correspond to the maximal chains in the lattice of flats of @xmath7 ; _ i.e. _ , the partition lattice @xmath288 .
thus we have given an alternative proof of the fact that there are @xmath289 maximal chains in @xmath288 ( * ? ? ?
3.3 ) . as an illustration of corollary [ totalfinecells ] , notice that , in figure [ assoc ] , the positive bergman complex @xmath247 consists of @xmath290 fine cells , while the bergman complex @xmath291 consists of @xmath292 fine cells . g. bergman . _ the logarithmic limit - set of an algebraic variety .
_ transactions of the american mathematical society , vol .
157 , 459 - 469 , 1971 .
l. billera , s. holmes , and k. vogtmann .
_ geometry of the space of phylogenetic trees .
_ advances in applied mathematics , vol .
27 , 733 - 767 , 2001 .
a. bjrner , m. las vergnas , b. sturmfels , n. white , g. ziegler .
_ oriented matroids .
_ encyclopedia of mathematics and its applications , vol . 46 .
cambridge university press , cambridge , 1993 .
m. einsiedler , m. kapranov , d. lind , and t. ward .
_ non - archimedean amoebas .
_ preprint , 2003 . g. lusztig .
_ introduction to total positivity _ , in positivity in lie theory : open problems , ed
. j. hilgert , j.d .
lawson , k.h .
neeb , e.b .
vinberg , de gruyter berlin , 1998 , 133 - 145 . | we study the positive bergman complex @xmath0 of an oriented matroid @xmath1 , which is a certain subcomplex of the bergman complex @xmath2 of the underlying unoriented matroid @xmath3 .
the positive bergman complex is defined so that given a linear ideal @xmath4 with associated oriented matroid @xmath5 , the positive tropical variety associated to @xmath4 is equal to the fan over @xmath6 .
our main result is that a certain fine " subdivision of @xmath0 is a geometric realization of the order complex of the proper part of the las vergnas face lattice of @xmath1 .
it follows that @xmath0 is homeomorphic to a sphere . for the oriented matroid of the complete graph @xmath7
, we show that the face poset of the coarse " subdivision of @xmath8 is dual to the face poset of the associahedron @xmath9 , and we give a formula for the number of fine cells within a coarse cell . |
the study of different algebraic structures using graph theory becomes an exciting research topic in the last few decades , leading to many fascinating results and questions .
given an algebraic structure @xmath10 , there are different formulations to associate a directed or undirected graph to @xmath10 , and we can study the algebraic properties of @xmath10 in terms of properties of associated graphs . directed power graphs associated to semigroups were introduced by kelarev and quinn @xcite .
if @xmath10 is a semigroup , then the directed power graph @xmath11 of @xmath10 is a directed graph with @xmath10 as the set of all vertices and for any two distinct vertices @xmath12 and @xmath13 of @xmath10 , there is an arc from @xmath12 to @xmath13 if @xmath14 for some positive integer @xmath15 .
then chakrabarty , ghosh and sen @xcite defined the undirected power graph @xmath16 of a semigroup @xmath10 as the underlying undirected graph of @xmath11 .
thus two distinct elements @xmath12 and @xmath13 of @xmath10 are edge connected in @xmath16 if @xmath17 or @xmath18 for some positive integer @xmath15 .
they proved that for a finite group @xmath3 , the undirected power graph @xmath19 is complete if and only if @xmath3 is a cyclic group of order @xmath20 or @xmath21 for some prime @xmath22 and positive integer @xmath15 . in @xcite
, cameron showed that for two finite commutative groups @xmath23 and @xmath24 , @xmath25 implies that @xmath26 .
they also observed that two finite groups with isomorphic undirected power graphs have the same number of elements of each order .
singh and manilal @xcite introduced strong power graph as a generalization of the undirected power graph of a finite group .
let @xmath3 be a group of @xmath27 elements .
the strong power graph @xmath9 of @xmath3 is a simple undirected graph whose vertices consists of the elements of @xmath3 and two distinct vertices @xmath4 and @xmath5 are adjacent in @xmath9 if @xmath28=@xmath29 for some positive integers @xmath30 .
thus a finite group @xmath3 is noncyclic if and only if @xmath9 of is complete .
also @xmath9 is connected if and only if @xmath27 is composite . here
we give several graph theoretic and spectral characterizations of the strong power graph of a finite group .
the strong power graph @xmath9 is a cayley graph of some group if and only if @xmath3 is noncyclic .
we give a complete list of the finite groups @xmath3 such that @xmath9 is a line graph .
such characterizations of the strong power graphs are given in section 2 . in section 3 , the algebraic connectivity and the chromatic number of @xmath9
have been found . for any graph @xmath31 ,
let @xmath32 be the adjacency matrix and @xmath33 be the diagonal of vertex degrees .
then the laplacian matrix of @xmath31 is defined as @xmath34=@xmath35 .
clearly @xmath34 is a real symmetric matrix and it is well known that @xmath36 is a positive semidefinite matrix with @xmath37 as the smallest eigen value .
thus we can assume that the laplacian eigen values are @xmath38 . among all laplacian eigen values of a graph ,
one of the most popular is the second smallest , called by fiedler @xcite , the algebraic connectivity of a graph .
it is a good parameter to measure , how well a graph is connected .
for example , a graph is connected if and only if its algebraic connectivity is non - zero . according to mohar @xcite the laplacian eigen values are more intuitive and much more important than the eigen values of the adjacency matrix .
we give a complete characterization of the laplacian spectrum of strong power graph of any finite group in section 4 . in section 5 , we have derived an explicit formula for the permanent of the laplacian matrix of strong power graphs of any finite group .
let @xmath39 be a square matrix of order @xmath27 , then the permanent of @xmath40 is denoted by per@xmath41 , where @xmath42 is the set of all permutations of @xmath43 .
the permanent function was introduced by binet and independently by cauchy in 1812 .
we refer to @xcite and @xcite for more on permanent .
in this section we characterize the groups @xmath3 such that the strong power graph @xmath9 is a line graph and cayley graph .
the line graph @xmath34 of a graph @xmath31 is constructed by taking the edge of @xmath31 as vertices of @xmath34 , and joining two vertices in @xmath34 whenever the corresponding edges in @xmath31 have a common vertex . a graph @xmath31 is called a line graph if it is a line graph of some graph .
the induced subgraph @xmath44 of a graph @xmath31 is a subgraph whose vertex set @xmath45 and all the edges whose endpoints are contained in @xmath46 .
the graph obtained by taking the union of graphs @xmath31 and @xmath47 with disjoint vertex set is the disjoined union or sum , denoted by @xmath48 .
the following result is a fundamental characterization of the simple line graphs .
proof of this result can be found in @xcite .
a simple graph @xmath3 is the line graph of some simple graph if and only if @xmath3 does not have any of the nine graphs below as an induced subgraph . for any positive integer @xmath27
, @xmath49 denotes the group of all integers modulo @xmath27 .
let @xmath50 .
thus @xmath51 for every prime @xmath22
. then we have @xmath52 @xmath53 or a prime . if @xmath54 is a line graph then @xmath55 or a prime .
if possible , on the contrary , suppose that @xmath56 and a prime .
then it follows that @xmath57 which implies that @xmath58 , and hence @xmath49 has at least three non - generators say , @xmath59 and @xmath60 . in this case
the following graph is an induced subgraph of @xmath54 .
therefore , by lemma 2.1 , @xmath54 can not be a line graph .
now we are ready to characterize all groups @xmath3 such that @xmath9 is a line graph .
let @xmath3 be a finite group of order @xmath27 . 1 . if @xmath3 is non - cyclic , then @xmath9 is a line graph of @xmath61 .
2 . if @xmath3 is cyclic , then @xmath9 is line graph if and only if @xmath55 or a prime .
if @xmath3 is non - cyclic then @xmath9 is a complete graph with @xmath27 vertices , and hence @xmath62 .
let @xmath3 be a cyclic group .
since strong power graphs of two isomorphic groups are isomorphic , so it is sufficient to prove the result for @xmath63 .
first suppose that @xmath55 or a prime .
we have @xmath64 is the line graph of the graph and @xmath65 is the line graph of the graph if @xmath66 , a prime number then @xmath67 is the line graph of @xmath68 .
converse , follows from the lemma 2.2 .
the notion of caley graph was first introduced by sir arthur cayley ( 1878 ) to produce a large class of natural example of vertex transitive graphs .
let @xmath3 be a finite group of order @xmath27 and @xmath10 be a subset of @xmath3 not containing @xmath69 , the identity of @xmath3 satisfying @xmath70 .
then the cayley graph of @xmath3 is denoted by @xmath71 is the graph whose vertex set is @xmath3 and two distinct vertices @xmath72 are edge connected if @xmath73 .
let @xmath3 be a group of order @xmath27 .
then @xmath9 is cayley graph of some group @xmath23 if and only if @xmath3 is noncyclic .
let @xmath3 is a noncyclic group of order @xmath27 .
then @xmath9 is a complete graph .
consider @xmath74 and @xmath75 , then clearly @xmath9 is cayley graph of @xmath74 .
if @xmath3 is a cyclic group of order @xmath27 , then @xmath9 is not regular . but
every cayley graph is regular .
hence @xmath9 is not a cayley graph of any group .
in this section we have discussed about vertex connectivity and chromatic number of the strong power graph @xmath9 of any finite group g. the * vertex connectivity or connectivity * of a graph @xmath76 is the minimum number of vertices in a vertex set @xmath10 such that @xmath31 @xmath77 @xmath10 is disconnected or has only vertex .
it is denoted by @xmath78.thus if @xmath31 is a complete graph with @xmath27 vertices then @xmath79=@xmath80 , and if @xmath31 is a disconnected graph then @xmath79=@xmath37 .
the strong power graph @xmath9 of every finite noncyclic group @xmath3 is complete .
thus we have : for any non cyclic finite group g of order n the vertex connectivity @xmath81 is n-1 .
let g be a cyclic group of order @xmath82 @xmath831 . *
if n is a prime number then @xmath81= @xmath37 . * if n is a composite number then @xmath84 = @xmath85 .
every non identity element of the group @xmath3is of order @xmath27,and no element of @xmath3 is adjacent with the identity element @xmath69 of the group @xmath3.thus @xmath9 is a disconnected graph with two components @xmath86 and @xmath3 @xmath77 @xmath86 , which implies that @xmath81=@xmath37 .
if @xmath27 is composite then @xmath85 elements of the finite cyclic group @xmath3 are non identity and non generators .
each of these non identity non generators is adjacent to every other vertex , which implies that to make the graph @xmath9 disconnected we have to remove at least these @xmath85 non identity non generators .
thus @xmath81 @xmath87 @xmath85 . since each generator is adjacent to every non identity element , so the removal of these @xmath85 non identity non generators makes the remaining graph disconnected with two components , one containing @xmath69 only and other containing all the generators .
thus @xmath81=@xmath85 .
the * chromatic number * of a graph @xmath31 is the minimum number of colours required to colour the vertices so that adjacent vertices get distinct colours .
let g be a cyclic group of order n , then the chromatic number @xmath88 of the strong power graph @xmath9 is n-1 .
the subgraph @xmath9 @xmath77 @xmath69 is complete . since the chromatic number of a complete graph with @xmath80 vertices is @xmath80 , so the chromatic number @xmath88 @xmath89 .
since the generators of the cyclic group @xmath3 are not adjacent with the identity @xmath69 of @xmath3 , either of of colours of the generators can be given to @xmath69 and hence @xmath88 @xmath90 .
thus @xmath88 = @xmath80 .
let g be a non - cyclic group of order n , then the chromatic number of the strong power graph @xmath9 is n. since the strong power graph @xmath9 of the cyclic group @xmath3 is complete , hence the result follows .
for each positive integer @xmath91 let @xmath92@xmath93 are non generators of @xmath49 then @xmath94 = @xmath95 , where m=@xmath85 . the laplacian matrix @xmath96 is the n times n matrix whose rows and columns are indexed in order by the non generators @xmath92@xmath93 and the generators of @xmath49 and @xmath97 is in last position .
then @xmath98 + each row and column sum of the above matrix is zero .
then the characteristic polynomial of @xmath96 is @xmath99= @xmath100 multiplying the first row of @xmath99 by @xmath101 and applying the row operation @xmath102=@xmath103 .
then the above determinant becomes @xmath104 again multiplying the first row of @xmath105 by @xmath106 and applying the row operation @xmath102=@xmath107 we get @xmath108 and so @xmath109 continuing in this way we get @xmath110 where @xmath111= @xmath112 @xmath113 , @xmath114 and order of the determinant is @xmath115 .
now expanding @xmath116 with respect to last row we get @xmath117 thus @xmath118 .
let g is a cyclic group of order n , then @xmath3 is isomorphic to @xmath49 ; and the strong power graphs @xmath9 and @xmath54 of @xmath3 and @xmath49 respectively are isomorphic .
hence the graphs @xmath9 and @xmath54 have the same laplacian spectrum .
so by theorem @xmath119 we have : if @xmath3 is a cyclic group of order @xmath27,then the laplacian spectrum of @xmath9 is @xmath120 for any non cyclic group @xmath3 , the strong power graph @xmath9 is complete @xcite .
so their laplacian spectrum is given by : let @xmath3 be a noncyclic group of order @xmath27 , then the laplacian spectrum of @xmath9 is @xmath121 the algebraic connectivity of a graph @xmath31 , denoted by @xmath122 , is the second smallest laplcian eigen value of @xmath31 @xcite .
now , the algebraic connectivity has received special attention due to its huge applications on connectivity problems , isoperimetric numbers , genus , combinatorial optimizations and many other problems .
thus by theorem @xmath119 and proposition @xmath123 we have : let @xmath3 be a group of order @xmath27 . 1 .
if @xmath3 is a cyclic group then @xmath124 .
2 . if @xmath3 is a noncyclic group then @xmath125 .
another important application of laplacian spectrum is on the number of spanning trees of a graph .
a spanning tree @xmath44 of a graph @xmath31 is a subgraph which is a tree having same vertex set is same as @xmath31 .
if @xmath126 @xmath87@xmath127@xmath87@xmath128@xmath87@xmath129@xmath87@xmath130=@xmath37 are the laplacian eigenvalues of a graph @xmath31 of @xmath27-vertices , then the number of spanning trees of @xmath31 is denoted by @xmath131 is given by @xmath132 [ theorem 4.11 ; @xcite ] .
thus from the proposition @xmath133 and proposition @xmath123 we have : let @xmath3 be a group of order @xmath27 . 1 .
if @xmath3 is a cyclic group then @xmath134 .
2 . if @xmath3 is a noncyclic group then @xmath135 .
the graph energy is defined in terms of the spectrum of the adjacency matrix .
depending on the well - developed spectral theory of the laplacian matrix , recently gutman et .
al @xcite have defined the laplacian energy of a graph @xmath31 with @xmath27 vertices and @xmath15 edges as : @xmath136 , where @xmath126 @xmath87@xmath127@xmath87@xmath128@xmath87@xmath129@xmath87@xmath130=@xmath37 are the laplacian eigen values of the graph @xmath31 .
this definition has been adjusted so that the laplcian energy becomes equal to the energy for any regular graph .
for various properties of laplacian energy we refer @xcite , @xcite , @xcite . from proposition
@xmath133 and proposition @xmath123 we have let @xmath3 be a finite group of order @xmath27 . 1 .
if @xmath3 is cyclic then @xmath137 .
if @xmath3 is noncyclic then @xmath138 .
let us recall the definition of permanent of a square matrix . for any square matrix @xmath139 of order @xmath27 ,
the permanent of a is denoted by per(a ) and per(a ) = @xmath140 .
it is quite difficult to determine the permanent of a square matrix . in this section
we have determined the permanent of the laplacian matrix of strong power graph of any finite group explicitly .
our method is based on the following observation .
let a = @xmath141 be a matrix of order @xmath27 , then per(a ) is equal to the coefficient of @xmath142 in the expression @xmath143 . throughout the rest of this section
we make the following convention : for any @xmath27-functions @xmath144 and for any @xmath27 variables @xmath145 , the coefficient of @xmath146 in a polynomial @xmath147 will be denoted by @xmath148 . here
we first find the permanent of adjacency matrix of any graph with @xmath149 vertices such that @xmath150 vertices form a clique and the rest vertex is adjacent with @xmath27 vertices .
we have the following lemma , @xmath157 @xmath158(\sum_{i=1}^{n}x_{m+i}).\\ & = n c_{x_1x_2 \cdots x_{m+n}}(x_{m+1}\prod_{i=1}^{m}(x - x_i)[\sum_{i=1}^{n}(x - x_{m+1})(x - x_{m+2})\cdots\widehat{(x - x_{m+i } ) } \cdots(x - x_{m+n})])\\ & = n c_{x_1x_2 \cdots x_m x_{m+2 } \cdots x_{m+n}}(\prod_{i=1}^{m}(x - x_i)[\sum_{i=1}^{n}(x - x_{m+1})(x - x_{m+2})\cdots\widehat{(x - x_{m+i } ) } \cdots(x - x_{m+n})])\\ & = n[c_{x_1x_2 \cdots x_m x_{m+2 } \cdots x_{m+n}}((x - x_1)\cdots ( x - x_m)(x - x_{m+2})\cdots ( x - x_{m+n}))+ \\ & \hspace{1.5 cm } c_{x_1x_2 \cdots x_m x_{m+2 } \cdots x_{m+n}}((x - x_1)\cdots ( x - x_m)(x - x_{m+1})(x - x_{m+3})\cdots ( x - x_{m+n}))+\cdots \\
\cdots & \hspace{3.5 cm } + c_{x_1x_2 \cdots x_m x_{m+2 } \cdots x_{m+n}}((x - x_1)\cdots ( x - x_m)(x - x_{m+1})\cdots ( x - x_{m+n-1}))]\end{aligned}\ ] ] now from the above lemma we get the permanent of adjacency matrix of strong power graph of any finite cyclic group .
let @xmath3 be a cyclic group of order @xmath27 .
then @xmath3 has @xmath167 generators none of which is adjacent to the identity element @xmath69 of @xmath3 .
thus the set @xmath168 of all @xmath80 non - identity vertices forms a clique and identity @xmath69 is adjacent to each of the @xmath169 non - identity non - generators .
hence from lemma @xmath170 it follows immediately that : the permanent of the laplcian matrix of a graph @xmath31 with @xmath149 vertices such that @xmath150 vertices form a clique and the rest vertex is edge connected with @xmath27 vertices is @xmath172 , where @xmath173 $ ] , and @xmath174 .
consider @xmath175 , where @xmath176 .
then the permanent of the laplacian matrix of @xmath31 is @xmath177 which is equal to @xmath178 .
now @xmath179 $ ] , where @xmath180 is a product of some @xmath181 and @xmath182 which is clear from the context .
so @xmath183 , where @xmath184 now proceeding in the proof of lemma @xmath170 we get @xmath185 $ ] .
s. chattopadhyay , p. panigrahi , on laplacian spectrum of power graphs of finite cyclic and dihedral groups , linear and multilinear algbra ( 2014 ) .
m. fielder , algebraic connectivity of graphs , czechoslovak math.j 23(1973 ) 298 - 305 .
y. hou , z. teng , c. w. woo on the spectral redius , @xmath192 degree and the upper bound of energy in a graph , match commun.math .
chem 57(2007)341 - 350 .
b. mohar , the laplacian spectrum of graphs , in : y. alavi , g. chartrand , oellermann or , a. j. schwenk , editors .
graph theory , combinatorics , and aplications , new york(nk):wiley , 2(1991 ) 871 - 898 . | let @xmath0 be a finite group of order @xmath1 .
the strong power graph @xmath2 of @xmath3 is the undirected graph whose vertices are the elements of @xmath3 such that two distinct vertices @xmath4 and @xmath5 are adjacent if @xmath6=@xmath7 for some positive integers @xmath8 . in this article
we classify all groups @xmath3 for which @xmath9 is line graph and caley graph .
spectrum and permanent of the laplacian matrix of the strong power graph @xmath9 are found for any finite group @xmath3 .
department of mathematics , visva - bharati , santiniketan-731235 , india .
+ [email protected] , [email protected] * keywords : * groups , strong power graphs , cayley graphs , laplacian spectrum , laplacian permanent . +
* 2010 mathematics subject classification : * 05c25 , 05c50 . |
rx j2115 - 58 was discovered as part of the _ rosat _ all - sky survey and independently by _ euve_. further observations by schwope et al ( 1997 ) and ramsay et al ( 1999 ) showed that it was an asynchronous polar with a beat period of 6.3 days .
such a short period allows us to study the effect of a changing orientation between the magnetic field and accretion stream on an easily observable timescale .
the left plot in fig . 1 shows phase - folded line profiles of h@xmath0 ( top ) , and heii 4686 ( bottom ) on the nights of 2000 july 20 , 22 and 24 , evenly sampling the beat cycle ( left to right ) .
these were taken using the aat rgo spectrograph .
the profiles are complex , with significant changes evident over the beat cycle .
an example is the enhanced heating of the secondary on the first night , evident in the strong narrow profile in the h@xmath0 line ( but not the heii 4686 ) , and the enhancement at @xmath1 on july 22 and 24 , which is manifest in a brightening of the ballistic stream in the doppler tomograms at these beat phases .
the plot on the right of fig .
1 shows light curves in hard x - rays from our _ xmm - newton _ observations of this system .
they consist of 7 separate observations of two orbital periods , one each day , evenly sampling the beat period .
they are extremely variable from night to night , but these in the hard band are not dissimilar to those seen in the _ rxte _ x - ray light curves ( ramsay et al 2000 ) .
we also have uv light curves from the _ xmm optical monitor_. accretion is mainly at one pole in the lower hemisphere at beat phase 0.87 , while it is clearly also accreting at a pole in the upper hemisphere at most other phases ( for example phase 0.23 ) .
the x - ray spectra also change appreciably at different beat phases : one ( at beat phase 0.13 ) requires a distinct soft x - ray component while the other ( at phase 0.55 ) does not .
similar behaviour is seen in our shorter xmm - newton observations of the asynchronous polar by cam ( ramsay & cropper 2002 ) .
a comprehensive analysis is in progress .
ramsay , g. , cropper , m. 2002 , mnras , 336 , 1129 ramsay , g. , buckley , d. a. h. , cropper , m. , harrop - allin , m.k . 1999 ,
mnras , 303 , 96 ramsay , g. , potter , s. , cropper , m. , buckley , d. a. h. , harrop - allin , m. k. 2000 , mnras , 316 , 225 schwope , a. , buckley , d. a. h. , odonoghue , d. , hasinger , g. , trmper , j. , voges , w. 1997 , a&a , 326 , 195 | we present here phase - resolved optical spectroscopy and x - ray light curves of the asynchronous polar rx j211558 as they change on the beat period .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
the gbar project aims at measuring the earth gravity acceleration hereafter denoted @xmath2 felt by an antihydrogen atom @xmath3 , using a free - fall technique at first@xcite and possibly spectroscopy of @xmath3 gravitational states in the future@xcite .
other collaborations aegis@xcite , athena - alpha@xcite , atrap@xcite pursue the same goal using different methods .
the specificity of the gbar project is to prepare a single antihydrogen atom at a temperature of the order of 10 @xmath4k , to obtain a sub - percent accuracy@xcite on @xmath2 . after a brief explanation of the scheme proposed for the gbar experiment
, we discuss in detail the trapping and sympathetic cooling of antihydrogen ions . in the gbar experimental scheme@xcite
, neutral antihydrogen is prepared by photodetachment of the excess positron of a sympathetically cooled @xmath0 trapped ion .
the photodetachment is performed using a pulsed laser .
the start time of the free - fall is the photodetachment pulse time .
the stop time corresponds to the annihilation of the @xmath3 on the detection plate .
obviously , reducing the velocity spread of @xmath0 atoms is indispensable for determining @xmath2 with high precision .
the @xmath0 ions are produced in two steps in a collision cell by sending kev antiprotons from the elena ring on a room temperature positronium cloud .
the first step produces antihydrogen atoms and the second one the @xmath0 ions following the reactions : @xmath5 the reaction cross sections have been evaluated by p. comini et al.@xcite , predicting that bunches of a few @xmath0 can be produced using state - of - the - art ps sources . since positronium is much lighter than @xmath6 , the @xmath0 energy distribution is linked to that of the @xmath6 bunch produced by the elena ring@xcite ( whose expected characteristics are 100 kev mean energy and a 4@xmath7 mm mrad emittance ) .
the @xmath6 bunch is decelerated to an energy of a few kev using a drift tube .
the kinetic energy spread of the @xmath6 bunch and hence of the @xmath0 ion bunch is about 300 ev corresponding to a temperature of 2.3@xmath8 k. the relative resolution on @xmath2 that can be obtained measuring the free - fall time of a single particle is given by@xcite @xmath9 where @xmath10 and @xmath11 are the position and velocity dispersions in the vertical direction . in the case of a quantum particle at the heisenberg limit , @xmath10 and @xmath11
are linked by the uncertainty relation @xmath12 where @xmath13 is the @xmath3 inertial mass .
an optimum resolution @xmath14 is obtained for @xmath15 leading to @xmath16 2.6 mm / s and @xmath17 assuming @xmath18 and @xmath19 m. the recoil due to the absorption of the 1.64 @xmath4 m detachment photon is @xmath20 cm / s and may be set in the horizontal plane so that its influence on the free - fall vanishes .
the recoil due to the excess energy can be made small using threshold detachment .
the associated energy is 0.3 m / s under the realistic assumption that the photon energy is 1 @xmath4ev above detachment threshold .
thus photodetachment prevents reaching the optimal free - fall conditions even if the initial ion were perfectly motionless .
our goal vertical velocity dispersion is @xmath21 m / s , eq .
[ eq_deltagsurg ] is then dominated by the second term leading to @xmath22 = 0.4 per detected atom .
a 1% resolution can be obtained by averaging on 1600 events . a transverse initial velocity of the order of 1 m / s is also required to avoid a too large detection area for the @xmath0 annihilation plates .
those velocities correspond to energies of the order of 5.2 nev or 120 @xmath4k per degree of freedom .
the @xmath0 cooling challenge is to bridge a 10 to 11 orders of magnitude gap on the ion temperature , going from the classical world of particle beam physics to the ultimate frontiers of quantum world .
indeed , if one considers the ground state of a quantum harmonic oscillator of mass @xmath13 = 1 a.u . and
angular frequency @xmath23 , the velocity spread is given by @xmath24 .
@xmath25 m / s leads to @xmath26 mhz .
this is the typical secular motion frequencies that are achieved in ion traps , showing that the gbar requirements can only be satisfied using ground state cooling techniques@xcite .
@xmath0 , antimatter equivalent of h@xmath27 , is extremely fragile against collisions with regular matter such that buffer gas cooling is not possible .
moreover , @xmath0 is a single electronic level atom that can not be directly laser cooled .
hence , we propose to use sympathetic cooling by the lighest laser cooled ion : @xmath28be@xmath1 .
since only a few @xmath0 ions are expected in each bunch , a nearly 100% capture efficiency is required . to that end , the gbar project will first use capture and doppler laser cooling step in a mm scale rf linear trap before transferring a single @xmath0 ion into a miniaturized trap ( called precision trap in the following ) to perform ground state cooling of a be@xmath1/@xmath0 ion pair . in sect . [ sec_capt_and_cooling ] ,
we discuss the capture and sympathetic cooling of a @xmath29 ion in a big be@xmath1 crystal . in sect .
[ sec_precision_trap ] , we discuss the separation of the cold @xmath0 ion and injection in the precision trap for ground state sideband sympathetic cooling before neutralization and @xmath3 release . the capture apparatus is depicted in fig .
[ fig_capt_trap ] .
it is made of a rf quadrupole guide and a biased segmented linear trap described below .
ion bunch .
@xmath30= 30 mm.,width=264 ] if the @xmath31 are not precooled below 300 ev , the kinetic energy spread of the @xmath0 ion bunch is large , and requires a deep trap .
the trapping depth of a linear trap is given by @xmath32 where q is the stability parameter and @xmath33 the applied voltage .
the @xmath34 parameter is inversely proportional to the ion mass and is typically chosen between 0.05 and 0.6 ( larger values lead to important rf heating of the ions ) . in order to safely trap both be@xmath1 and @xmath0 , @xmath34 must be chosen close to 0.45 for @xmath0 and 0.05 for be@xmath1 .
we assume a trapping depth of 20 ev for @xmath0 that is obtained using @xmath35 356 v ( 712 v peak to peak ) .
the stability parameter of a linear trap is given by @xmath36 where @xmath13 and @xmath37 are the ion mass and charge , @xmath38 is the inner trap radius , and @xmath39 the rf frequency . with @xmath40
mm , we get @xmath41 mhz , i.e. , standard trap parameters .
efficient capture of @xmath0 ions with a 300 ev kinetic energy spread is much more involved , requiring 10 800 v peak to peak at 68.5 mhz .
the use of rf traps with 2 drive frequencies@xcite was envisaged but it requires 2 - 3 order of magnitude different mass - to - charge ratios .
the incoming @xmath0 ion bunch has 1 to 6 kev kinetic energy .
the ion bunch is decelerated by biasing the linear trap by @xmath42 980 to 5980 v. the trap input endcap voltage is lowered to @xmath43 for a short time @xmath44 for the @xmath0 ion bunch intake .
since the be@xmath1 ion motion is strongly damped by the cooling laser , and @xmath44 is much shorter than the axial trap secular period , the be@xmath1 ions do nt have time to escape the trap .
figure [ fig_capt_eff_tau_ener]-a shows a simulation of the capture efficiency ( without be@xmath1 ) versus the time delay @xmath44 . 100% capture efficiency is predicted for a large range of @xmath44 for a small kinetic energy spread @xmath45 ev .
figure [ fig_capt_eff_tau_ener]-b shows the capture efficiency for optimal intake time @xmath44 versus the @xmath0 bunch kinetic energy spread @xmath46 .
the efficiency decreases with @xmath46 , but remains larger than 50% for @xmath47 25 ev .
this analysis shows that the initial kinetic energy spread of antiprotons from the elena source has to be reduced by at least one order of magnitude to allow for their efficient capture with reasonable trap parameters .
capture efficiency versus trap opening time @xmath44 for different ion bunch durations , for @xmath45 v. ( b ) : @xmath0 capture efficiency versus kinetic energy spread for @xmath48s .
each point corresponds to a different simulation with 100 ions.,title="fig:",width=177 ] capture efficiency versus trap opening time @xmath44 for different ion bunch durations , for @xmath45 v. ( b ) : @xmath0 capture efficiency versus kinetic energy spread for @xmath48s .
each point corresponds to a different simulation with 100 ions.,title="fig:",width=177 ] once captured , @xmath0 ions have a very high temperature , limited by the trap depth .
next , they are sympathetically cooled by coulomb interaction with a large laser cooled be@xmath1 ion cloud .
because of possible photodetachment by the cooling laser light ( see sect . [ sec_photodetach ] ) , it is very important to evaluate the cooling time .
sympathetic cooling dynamics results from the competition between the coulomb repulsion and the trapping forces that take the ions together , and between laser cooling and rf heating . for this reason
, it can only be evaluated using ion dynamics numerical simulation taking into account the exact time - dependent trapping forces responsible for micromotion and rf heating , and the exact coulomb repulsion@xcite .
short time steps ( sub - ns range ) must be used to well represent the fast dynamics due to the rf field and secure calculation convergence , since long simulation times are required to get sympathetic cooling evidence .
the main numerical complication comes from the evaluation of the full coulomb interaction for a large number of ions .
the numerical simulations are done using a home built fortran code to solve the newton s equation of motion for @xmath49 laser cooled be@xmath1 ions and @xmath50 sympathetically cooled ions , whose masses were taken equal to 1 ( @xmath0 ) , 2 ( h@xmath51 ) and 3 ( hd@xmath1 ) in order to study the mass dependence of the cooling process . the equations are integrated using either a fixed - step fourth - order runge - kutta method or the leap - frog ( verlet - velocity)@xcite algorithm . the code takes into account the time - dependent rf trapping field and the axial harmonic trapping field of an ideal linear trap model given by the gradient of the potential @xmath52 where @xmath53 is the rf voltage , @xmath39 the rf angular frequency , @xmath38 is the effective inner radius of the ion trap , @xmath54 is the mass of the considered ion and @xmath55 its axial oscillation frequency . for two different ionic species labeled @xmath56 and @xmath57
, we have @xmath58 . the coulomb force undergone by ion @xmath56 is given by @xmath59 the laser action
is taken into account in terms of absorption , spontaneous and stimulated emission processes for a two - level atom in a gaussian laser beam of waist @xmath60 and wave vector @xmath61 . at each time step and for each laser - cooled ion , depending on its internal state and position in the laser beam , the absorption and emission probabilities are evaluated in a quantum jump approach . in case of absorption , stimulated or spontaneous emission , the ion velocity is changed by @xmath62 or by @xmath63 where @xmath64 is a random direction . at each time step , the ion positions are checked to be within a cylinder of radius @xmath38 and length @xmath30 or are withdrawn from the simulation .
the performance of the code is limited by the coulomb interaction evaluation , so the computation time scales as the square of the ion number .
the double precision coulomb force subroutine evaluates @xmath65 coulomb terms per second on a 3 ghz cpu .
using multi - core cpu s , we observe a proportional speed - up .
a @xmath49 laser cooled be@xmath1 ion cloud is numerically prepared and relaxed to equilibrium , and @xmath50 @xmath0 or h@xmath51 ions are introduced along the trap axis , next to the be@xmath1 ion cloud ( see fig .
[ fig_ion_cloud ] ) corresponding to a potential energy of a few mev .
one can distinguish two cooling phases . at the beginning of the cooling process
, the sympathetically cooled ion goes in and out the be@xmath1 ion cloud , and only periodically interacts with the coolant ions , progressively losing secular kinetic energy . during this first phase
, the be@xmath1 ion cloud is not crystallized .
once the sympathetically cooled ion gets embedded in the be@xmath1 cloud , a more efficient cooling phase then starts finally leading to a mixed species coulomb crystal as shown in the left part of fig .
[ fig_ion_cloud ] , which illustrates the second phase of the sympathetic cooling dynamics for a single h@xmath51 or @xmath0 ion by 2000 be@xmath1 laser - cooled ions . and 1 h@xmath51 ion cloud .
the sympathetically cooled h@xmath51 ion is the red circle .
the purple arrow shows the cooling laser direction .
the cooling laser is aligned on the trap axis @xmath66 with a waist @xmath67 m located at the trap center .
the laser detuning is @xmath68 and the laser intensity on the axis is @xmath69 .
* graphs * : macromotion kinetic energy for each degree of freedom , averaged over 170 rf periods for 2000 laser cooled be@xmath1 ions ( a , c ) and for one h@xmath51 ( b ) or @xmath0 ( d ) sympathetically cooled ion .
black circles : @xmath70 , red squares : @xmath71 , green diamonds : @xmath66 .
the horizontal dashed line corresponds to the doppler cooling limit temperature @xmath72 leading to @xmath73 j. the dash - dotted line in ( b ) corresponds to a 1 ms exponential decay behavior .
for all the graphs , the trap parameters appearing in eq .
( 3 ) are : @xmath74 mm , @xmath75 v , and @xmath76 mhz .
the integration time step is @xmath77 s. ( a ) and ( b ) : @xmath78 v , @xmath79 khz for @xmath80 .
( c ) and ( d ) : @xmath81 v , @xmath82 khz for @xmath80.,title="fig:",width=136 ] and 1 h@xmath51 ion cloud . the sympathetically cooled h@xmath51 ion is the red circle .
the purple arrow shows the cooling laser direction .
the cooling laser is aligned on the trap axis @xmath66 with a waist @xmath67 m located at the trap center .
the laser detuning is @xmath68 and the laser intensity on the axis is @xmath69 .
* graphs * : macromotion kinetic energy for each degree of freedom , averaged over 170 rf periods for 2000 laser cooled be@xmath1 ions ( a , c ) and for one h@xmath51 ( b ) or @xmath0 ( d ) sympathetically cooled ion .
black circles : @xmath70 , red squares : @xmath71 , green diamonds : @xmath66 .
the horizontal dashed line corresponds to the doppler cooling limit temperature @xmath72 leading to @xmath73 j. the dash - dotted line in ( b ) corresponds to a 1 ms exponential decay behavior .
for all the graphs , the trap parameters appearing in eq .
( 3 ) are : @xmath74 mm , @xmath75 v , and @xmath76 mhz .
the integration time step is @xmath77 s. ( a ) and ( b ) : @xmath78 v , @xmath79 khz for @xmath80 .
( c ) and ( d ) : @xmath81 v , @xmath82 khz for @xmath80.,title="fig:",width=340 ] we plot the averaged macromotion kinetic energy in the @xmath70 , @xmath71 and @xmath66 directions ( see figure [ fig_ion_cloud ] caption ) . in the case of h@xmath51
, we observe an exponential decay of the transverse kinetic energies down to the doppler limit with a time constant of 1 ms , and a much faster decay of the axial kinetic energy , indicating the feasibility of fast sympathetic cooling for a 9/2 ion mass ratio .
for @xmath0 , the situation is quite different .
whereas the axial motion is quickly damped to the doppler limit , the competition between rf heating and sympathetic cooling in the transverse direction leads to a high transverse @xmath0 kinetic energy corresponding to temperatures in the k range .
indeed , the motional coupling between two particles of different masses rapidly decreases with the mass ratio .
it is thus important to work out more efficient sympathetic cooling schemes .
one solution is to use an intermediate mass ion@xcite such as hd@xmath1 with a mass of 3 . the left part of fig .
[ fig_cooling_be_hd_h ] shows that starting with a coulomb crystal made of 1800 be@xmath1 and 200 hd@xmath1 ions , ms @xmath0 cooling times are achievable .
/ 200 hd@xmath1/ 1 @xmath0 mixed ion cloud .
numerical parameters as in fig .
[ fig_ion_cloud](d ) .
the @xmath0 initial position is on the trap axis , 0.5 mm from the center , corresponding to a 4.6 mev potential energy . *
right * : doppler cooling dynamics of a be@xmath1/x@xmath1 ion pair in the precision trap .
the mass of the sympathetically cooled ion corresponds either to real ( black @xmath80 , green @xmath83 ) or fictitious ion masses ( blue @xmath84 , red @xmath85 ) . ( d ) 3d kinetic energy , ( e ) kinetic energy of the @xmath66 motion.,title="fig:",width=245 ] / 200 hd@xmath1/ 1 @xmath0 mixed ion cloud .
numerical parameters as in fig .
[ fig_ion_cloud](d ) .
the @xmath0 initial position is on the trap axis , 0.5 mm from the center , corresponding to a 4.6 mev potential energy . *
right * : doppler cooling dynamics of a be@xmath1/x@xmath1 ion pair in the precision trap .
the mass of the sympathetically cooled ion corresponds either to real ( black @xmath80 , green @xmath83 ) or fictitious ion masses ( blue @xmath84 , red @xmath85 ) .
( d ) 3d kinetic energy , ( e ) kinetic energy of the @xmath66 motion.,title="fig:",width=245 ] the photodetachment threshold of @xmath0 ( 1.64 @xmath4 m ) is well below the 313 nm be@xmath1 cooling photon energy , so the cooling laser beam may photodetach the excess positron . for a be@xmath1 cooling beam at saturation intensity ( @xmath86=0.82 mw / mm@xmath87 with @xmath88 mhz ) , the photon flux is @xmath89 photon / s / cm@xmath87 .
the photodetachment cross section of h@xmath27 at 313 nm@xcite is @xmath91 cm@xmath87 , leading to @xmath92 s@xmath93 . under those conditions ,
the @xmath0 lifetime is less than 1s .
it can be made longer using a lower cooling intensity or by using a quasi continuous cooling beam with a reduced duty cycle . in a rf trap ,
the trapping effective potential is tighter for light ions than for heavy ones .
taking advantage of this fact , one might use a hollow laser beam in a gauss - laguerre mode@xcite l@xmath94 to cool the be@xmath1 while the @xmath0 ions which are strongly confined very close to the trap axis are exposed to a negligible amount of laser radiation at 313 nm .
as mentioned in the introduction , the doppler limit temperature is not low enough for the gbar project , so the @xmath0 ion will be injected in a precision trap to form a be@xmath1/@xmath0 ion pair on which ground state raman side band cooling can be performed @xcite .
the precision trap ( see fig [ fig_raman_transition_precision_trap])@xcite consists of four gold coated , micro - fabricated alumina chips which are arranged in an x - shaped configuration and two endcaps made from titanium . the endcaps are pierced with a hole with a diameter of 600 @xmath4 m to enable ion injection into the trap .
two of the chips provide the rf - field and two the dc trapping potential respectively .
the chips have 11 electrodes each to shape the axial potential what for the voltages can be controlled with a custom built digital - to - analog converter with a voltage resolution of 300 @xmath4v and a time resolution of 400 ns .
the distance between the chips is 960 @xmath4 m .
the trap is driven by an rf - voltage with a frequency @xmath95 and a peak - to - peak amplitude @xmath96 .
this leads to a q parameter for @xmath97 of 0.05 and for @xmath29 of 0.45 .
the center dc - electrode is held at -1.5 v to provide axial confinement .
this voltage configuration leads to a axial and radial secular frequencies @xmath98 and @xmath99 for @xmath0 .
the corresponding ground state kinetic energies are 0.09 and 0.41 mk .
after extraction from the capture trap , a cold @xmath0 ion is injected into the precision trap through the end cap
. the right part of fig .
[ fig_cooling_be_hd_h ] shows sympathetic doppler cooling of a be@xmath1/x@xmath1 ion pair where the precision trap is modeled as an ideal linear paul trap .
the cooling time strongly depends on the x@xmath1 mass and can be larger than seconds in the case of @xmath0 . again
, the z - motion cooling is much faster due to the absence of rf heating . in the precision trap , the cold be@xmath1 and @xmath0 ions
are coupled harmonic oscillators . the doppler limit temperature ( 0.47 mk ) corresponds to excitations of a few quanta that can be further laser cooled .
the motional couplings of an ion pair confined in a rf paul trap have been evaluated by wbbena et al.@xcite . in the @xmath70 , @xmath71 or @xmath66 directions ,
the individual ion trajectories can be expanded on _ in - phase _ and _ out - of - phase _ eigenmodes as @xmath100 where the amplitudes @xmath101 and @xmath102 and phases @xmath103 and
@xmath104 depend on the initial conditions .
the motional coupling coefficients @xmath105 and @xmath106 depend on the particle masses and trapping conditions and verify @xmath107 ( see eq .
( 14 ) and ( 17 ) in@xcite ) . for the axial motion , we get @xmath108 and @xmath109 . for the transverse motion , assuming @xmath110 , we get @xmath111 and @xmath112 , which can be compared to @xmath113 in the case of an ion pair with equal masses , or 1 for a single ion
. figure [ fig_raman_transition_precision_trap ] shows the be@xmath1 electronic energy levels with the confined ion vibrational structure .
raman side band cooling consists in using an off - resonance stimulated raman transition and a resonant spontaneous raman transition to decrease the vibration number down to 0 .
a stimulated raman transition is a coherent process .
the time required to drive a @xmath7-pulse is inversely proportional to the dipole matrix element , i.e. to the coupling coefficient @xmath106 . for a be@xmath1/@xmath0 ion pair ,
it is at most 60 times longer than for a single ion . in the latter case ,
raman sideband cooling to the ground state for the three degree of freedom can be performed within a few tens of ms@xcite . for a be@xmath1/@xmath0 ion pair
, it may thus be achieved for the 6 degrees of freedom within 1 s. eigenmodes with frequencies in the 2 - 8 mhz range make it possible to efficiently laser cool the ion pair to its vibrational ground state .
the corresponding velocity dispersion is still slightly too large for the gbar experiment .
we thus propose to adiabatically ramp down the trapping stiffness down to @xmath114 30 khz radial and longitudinal oscillation frequencies by slowly lowering the trapping voltage within about 500 ms .
30 khz oscillation frequencies correspond to @xmath115 8 cm / s velocity dispersions which amply satisfies the gbar requirements .
we have discussed the challenges and feasibility of antihydrogen positive ion @xmath0 sympathetic cooling down to @xmath4k temperatures for the gbar project .
we here summarize the results .
* capture efficiency of a 1 kev @xmath0 ion bunch in a biased linear trap can be larger than 50% with a 25 ev kinetic energy dispersion and a trap depth of 20 ev . the efficient capture of an ion bunch with 300 ev dispersion requires out - of - reach trapping conditions , indicating that the antiprotons delivered by the elena ring have to be cooled beforehand to decrease the energy dispersion by at least a factor of 10 .
* sympathetic doppler cooling by laser cooled be@xmath1 ions is shown to be efficient if the ions remain embedded in the be@xmath1 ion cloud .
this means that one has to use large be@xmath1 ion clouds filling the capture trap .
the numerical simulations show that the cooling efficiency is much better with a 9/2 rather than with a 9/1 mass ratio . in the latter case
, the efficiency is dramatically improved using a third species of mass 3 , i.e. hd@xmath1 ions .
a possible experimental scheme is then to prepare a laser cooled be@xmath1 ion cloud and to inject hd@xmath1 ions from an external ion source ( in order not to increase the pressure in the vacuum chamber ) before the @xmath0 bunch intake .
sympathetic doppler cooling of energetic @xmath0 is the most challenging step and is still an open problem , which has to be tackled both experimentally and using numerical simulation . to that end and in order to perform numerical simulations of the cooling dynamics with large number of ions ( @xmath116 10000 ) , the code will be implemented on massively parallel graphic processing units ( gpu ) . from the experimental point of view , this step will be first tested using matter ions h@xmath51 and h@xmath1 ( protons ) . *
the transfer of a single @xmath0 ion from the capture trap to the precision trap , which was not discussed here , will be done using standard ion beam optics for injection through the drilled end cap .
the experimental protocol will be worked out first with matter ions with ca@xmath1/be@xmath1 and then be@xmath1/h@xmath51 and be@xmath1/protons before being implemented on gbar . here , the main issue is to avoid heating the ion during the transfer to secure a fast re - capture and sympathetic doppler cooling of the ion pair .
* we have shown that once a be@xmath1/@xmath0 ion pair is prepared at the doppler limit temperature , raman sideband sympathetic cooling down to the vibrational ground state of the trap is feasible with less than one second , preparing a @xmath0 ion with a velocity dispersion of about 1 m / s . *
the velocity dispersion can be decreased to about 10 cm / s by adiabatically ramping down the trapping stiffness by a factor of 100 , within less than 0.5 s. at that point , the velocity dispersion of the antihydrogen produced by threshold photodetachment of the excess positron is dominated by the recoil due to the e@xmath1 ejection .
we thank jofre pedregosa for helpul discussions and dominique delande for introducing us to open mp parallelization .
we also thank the cost action mp1001-iota , nanok - ifraf resima grant and anr / dfg anr-13-is04 - 0002 - 01 bescool grant .
00 j. walz and t. hnsch , _ general relativity and gravitation _ * 36 * , 561 ( 2004 ) .
p. perez and y. sacquin .
_ classical and quantum gravity _ * 29 * , 18 ( 2012 ) .
spsc - p-342 , proposal to measure the gravitational behaviour of antihydrogen at rest , gbar .
cern ( 2011 ) .
voronin , v. v. nesvizhevsky , and s. reynaud , _ phys .
* 85 * , 014902 ( 2012 ) .
g. dufour , a. grardin , r. gurout , a. lambrecht , v. v. nesvizhevsky , s. reynaud , and a. yu .
voronin , _ phys .
a _ * 87 * , 012901 ( 2013 ) .
a. kellerbauer et al . , _ nuclear instruments and methods in physics research section b _ * 266 * , 351 ( 2008 ) . the alpha collaboration and a. e. charman . _ nature communications _ * 4 * , 1785 ( 2013 ) .
g gabrielse .
_ technical report _
cern-spsc-2010-006.spsc-sr-057 , cern , geneva , jan 2010 .
gabriel dufour , pascal debu , astrid lambrecht , valery nesvizhevsky , serge reynaud and alexei voronin , to be published in the european physical journal c ( 2014 ) .
doi : 10.1140/epjc / s10052 - 014 - 2731 - 8 p. comini and p .- a .
hervieux , _ new journal of physics _ * 15 * , 095022 ( 2013 ) .
spsc - p-338 , elena : an upgrade to the antiproton decelerator
d. leibfried , r. blatt , c. monroe , d. wineland , _ reviews of modern physics _ * 75 * , 281 ( 2003 ) .
d. trypogeorgos , c. foot , http://arxiv.org/abs/1310.6294 .
k. lmmerzahl , s. schiller , _ physical review a _ * 68 * , 053406 ( 2003 ) .
l. verlet , _ physical review _ * 159 * , 98 ( 1957 ) . c. b. zhang , d. offenberg , b. roth , m. a. wilson , and s. schiller , _ phys .
rev . a _ * 76 * , 012719 ( 2007 ) .
s. chandrasekhar , d. d. elbert , _ astrophysical journal _ * 128 * , 633 ( 1958 ) . s. j. smith , d. s. burch , _ physical review letters _ * 2 * , 165 ( 1959 ) . c. t. schmiegelow , f. schmidt - kaler , european journal of physics d 66 , 157 ( 2012 ) . j. b. wbbena , s. amairi , o. mandel , p. o. schmidt , _ phys .
rev . a _ * 85 * , 043412 ( 2012 ) .
t. rosenband , p. o. schmidt , d. b. hume , w. m. itano , t. m. fortier , j. e. stalnaker , k. kim , s. a. diddams , j. c. j. koelemeij , j. c. bergquist , and d. j. wineland _ physical review letters _ * 98 * , 220801 ( 2007 ) . c. w. chou , d. b. hume , j. c. j. koelemeij , d. j. wineland , and t. rosenband _ physical review letters _ * 104 * , 070802 ( 2010 ) .
w. schnitzler , n. m. linke , r. fickler , j. meijer , f. schmidt - kaler and k. singer , _ physical review letters _ * 102 * , 070501 ( 2009 ) . | we discuss an experimental approach allowing to prepare antihydrogen atoms for the gbar experiment .
we study the feasibility of all necessary experimental steps : the capture of incoming @xmath0 ions at kev energies in a deep linear rf trap , sympathetic cooling by laser cooled be@xmath1 ions , transfer to a miniaturized trap and raman sideband cooling of an ion pair to the motional ground state , and further reducing the momentum of the wavepacket by adiabatic opening of the trap . for each step ,
we point out the experimental challenges and discuss the efficiency and characteristic times , showing that capture and cooling are possible within a few seconds . |
string theory has offered significant insights in answering many puzzles surrounding properties of black holes . arguably , one of the greatest successes of string theory is the statistical mechanical explanation of the bekenstein - hawking entropy of certain extremal and near - extremal black holes .
the first example was the five - dimensional black hole studied in @xcite .
following @xcite , a number of different types of black holes have been studied and the agreement between the statistical mechanical entropy and the bekenstein - hawking entropy has been shown to hold in each case ( see e.g. reviews @xcite and references therein ) . in particular , references @xcite extended the success of @xcite to the five - dimensional near - extremal setting .
the black holes considered in @xcite all have topologically spherical horizons . in 2004 ,
elvang , emparan , mateos , and reall ( eemr ) @xcite presented the first example of a supersymmetric solution with horizon topology @xmath8 , that is , a supersymmetric black ring .
the bekenstein - hawking entropy of the supersymmetric black ring was reproduced from a statistical mechanical description in @xcite using the maldacena - strominger - witten ( msw ) cft @xcite .
although , the application of msw cft gives an impressive match of the statistical and bekenstein - hawking entropies , it leaves many questions unanswered . for a critical discussion and further references
see @xcite .
it is hoped ( see e.g. @xcite ) that the matching of the statistical and bekenstein - hawking entropies for black rings in the thermally excited ( near - extremal ) setting will clarify many questions left unanswered in the supersymmetric limit .
a microscopic discussion for the entropy of the yet - to - be found black ring that describes thermal excitations over the supersymmetric ring of @xcite is already proposed in the literature @xcite . on the supergravity side ,
the construction of a black ring solution that describes thermal excitations over the supersymmetric black ring is a significantly involved problem . unlike the case of
topologically spherical black holes @xcite , one can not add a sufficient number of charges on a neutral black ring by applying familiar solution generating techniques such as boosts and string dualities .
for example , one can not add @xcite three independent m2 charges to an otherwise neutral ring by applying boosts and string dualities . adding the third m2 charge
typically requires applying a boost along the direction of a kk - monopole in an intermediate step .
such a boost is incompatible with the identifications imposed by the kk - monopole fibration . in effect
, one ends up generating pathological spacetimes .
a variant of this problem also arises in the straight string limit when one tries to add three m2 charges on top of a black string carrying three m5 charges @xcite .
this and related problems have resisted various attempts to construct a variety of non - extremal black string and black ring solutions @xcite .
reference @xcite shed some light on this problem from a group theoretic perspective in the simplified setting of minimal five - dimensional supergravity .
there the pseudo - compact generators of the three - dimensional hidden symmetry group were associated to charging transformations .
a proposal was made to construct the solution in the straight string limit by acting with several charging transformations .
the proposal however did not allow the authors of @xcite to write the solution in a manageable form .
in this paper we extend the investigations of @xcite and present a solution of this problem in minimal supergravity in the straight string limit , i.e. , we explicitly construct the most general black string solution of five dimensional minimal supergravity .
our black string has five independent parameters , namely , magnetic one - brane charge , smeared electric zero - brane charge , boost along the string direction , energy above the bps bound , and rotation in the transverse space . in our solution one is allowed to vary all five parameters independently . upon setting the magnetic or electric charge to zero ,
our solution reduces to the spinning non - extremal string solutions of @xcite .
our solution admits a four parameter family of extremal black strings corresponding to having maximal rotation in the transverse space .
we can recover the two known extremal black strings by taking additional limits . in one limit we recover the three parameter supersymmetric string of @xcite . in another extremal limit our string solution reduces to the three parameter non - rotating non - supersymmetric string of @xcite .
when the pressureless condition is imposed , our solution describes the infinite radius limit of the most general class of black rings of minimal supergravity . on the microscopics side
, the validity of the entropy formula in @xcite depends on the absence of finite ring radius corrections , which is not guaranteed _ a priori_. in the straight string limit this subtlety goes away .
so , it is expected that the entropy formula in @xcite reproduces the bekenstein - hawking entropy of our string in the near - extremal limit .
this indeed turns out to be the case once one has made precise the connection between quantities defined near the horizon and at infinity .
an important open issue in the black ring literature has been why various black ring solutions must obey a number of bounds on various parameters in order to be smooth .
for example , for the supersymmetric ring of minimal supergravity , the electric charge is bounded from below @xcite .
the lower bound is in terms of the dipole charge and the radius of the ring @xcite .
curiously enough , the bound persists in the infinite radius limit . although , we do not offer a complete understanding of these issues , our string solution does shed light on some limit of these puzzles as well .
the extremal pressureless limits of our non - rotating string provide us with a working phase diagram for a class of extremal black rings of minimal supergravity .
the phase diagram suggests that as the above mentioned bound on the electric charge of the supersymmetric black ring is violated , a non - supersymmetric branch of extremal black rings emerges .
this non - supersymmetric branch of extremal black rings has not yet been constructed in full generality . on this branch
one can continuously take the electric charge to zero , thus the non - supersymmetric branch of black rings is connected to a dipole ring of @xcite .
the dipole ring is such that in the infinite radius limit the boost along the string completely breaks supersymmetry .
our phase diagram in addition suggests that there is another branch of extremal black rings in minimal supergravity .
this branch is also connected to a dipole ring of @xcite .
the dipole ring on this branch is such that in the infinite radius limit the boost along the string _ preserves _ supersymmetry .
extremal black rings on this branch are also not known in full generality .
the rest of the paper is organized as follows . in section [ sec : setup ]
we introduce our notation and conventions . in section
[ sec : construction ] we construct the most general black string solution and analyze its physical properties and thermodynamics .
first , in section [ charge_matrix ] , we construct a group element of g@xmath9 in order to build an appropriate ` charge matrix ; ' for the benefit of the reader the concept of the ` charge matrix ' is reviewed .
next , in section [ sec : monster ] , we act on the kerr string with this group element to construct the most general string . physical properties and thermodynamics of our general solution
are presented in section [ thermo ] . in section [ sec :
limits ] we explicitly find the extremal limits of our solution . in [ extremal_limits ]
we show there are three distinct extremal limits , corresponding to a four - parameter non - supersymmetric string with maximal rotation in the transverse space , a three - parameter supersymmetric string , and a three - parameter non - rotating non - supersymmetric string . in [ electric ] and [ magnetic ] we briefly discuss how our solution reduces to the spinning @xmath10 and @xmath11 strings respectively . in section [ sec : decoupling ] we match the entropy of our general solution with the prediction of @xcite . in section [ decoupling ]
we outline how to properly take the decoupling limit and verify that our solution in this limit is @xmath12 . in section
[ matching ] we match the entropy with the cft prediction .
we finish in section [ sec : discussion ] with a summary and a discussion about a few implications that the work presented herein has for yet - to - be found black rings .
since we are interested in constructing a black string solution , we can compactify minimal five - dimensional supergravity along the string direction and construct the corresponding black hole in the resulting four - dimensional theory . this way to approaching
the problem turns out to be much more systematic , as there are a number of theorems @xcite to bank on concerning black holes in the resulting four - dimensional theory .
the resulting four - dimensional theory is a single modulus @xmath13 supergravity .
it is sometimes also called the s@xmath14 model ( or the @xmath15 model ) as its prepotential is cubic in the modulus .
it can be obtained by setting s@xmath16t@xmath16u in the stu model @xcite . in this section
we review salient features of @xmath13 single modulus s@xmath14 model .
we define electromagnetic charges in this model and discuss the uplift back to minimal five - dimensional supergravity .
let us start from the bosonic sector of eleven - dimensional supergravity @xmath18 where @xmath19 is a three - form and @xmath20 is its field - strength .
upon compactification on a six - torus with the ansatz ds^2_11 & = & ds^2_5 + dx_1 ^ 2 + dx^2_2 + + dx_6 ^ 2 + a & = & a ( d x_1 d x_2+dx_3 dx_4 + dx_5 dx_6 ) we obtain the bosonic sector of @xmath21 ( i.e. , minimal ) five - dimensional supergravity .
the lagrangian takes the form of einstein - maxwell theory with a chern - simons term , @xmath22 to perform the dimensional reduction to four dimensions , we assume that the extra spatial direction ( denoted @xmath23 ) is compact and is a killing vector in the five - dimensional spacetime . using the standard kaluza - klein ansatz to yield a four - dimensional lagrangian in einstein frame
we write the five - dimensional metric as @xmath24 from this ansatz one finds ( see for example @xcite ) that the resulting four - dimensional lagrangian takes the form @xmath25 where @xmath26 the scalars @xmath27 and @xmath28 parameterize an @xmath29 coset .
the four electromagnetic charges corresponding to the two vectors are defined as follows .
the equations of motion for the potentials @xmath30 and @xmath31 are @xmath32 and @xmath33 using , one can rewrite as the closure of a form @xmath34 , @xmath35 we then use the closed forms @xmath34 and @xmath36 given by and to define conserved electric charges in asymptotically flat four - dimensional spacetimes as integrals over a two - sphere at spatial infinity @xmath37 @xmath38 in a similar fashion , from the bianchi identities @xmath39 for @xmath30 and @xmath31 we define the magnetic charges @xmath40 note the minus sign in the definition of @xmath41 .
we work with sign conventions in which the static extremal black hole carrying positive @xmath41 and @xmath42 charges is supersymmetric . from the m - theory point of view
the electromagnetic charges correspond to the brane charges as indicated in table [ tablecharges ] : 0.2 cm 0.2 cm from table [ pressureless ] we can immediately draw a number of conclusions .
clearly , only case @xmath43 saturates the five - dimensional bps bound .
it corresponds to the black string of @xcite , which in turn corresponds to the infinite radius limit of the supersymmetric black ring of @xcite . in case
@xmath43 the m2@xmath14 electric charge @xmath44 is bounded from below @xmath45 . as this bound is violated , solution @xmath43 ceases to be smooth .
precisely when solution @xmath43 becomes singular , the pressureless non - supersymmetric solution , case @xmath46 , takes over .
see figure [ fig : phase ] .
branch @xmath46 is completely smooth . on this branch
one can continuously take the electric charge to zero .
thus the corresponding black ring is connected to an extremal dipole ring of @xcite .
the dipole ring is such that in the infinite radius limit , the boost along the string completely breaks supersymmetry .
black ring solutions on this branch are not known in full generality .
case @xmath47 , on the other hand , indicates that there is another branch of extremal black rings of minimal supergravity .
this branch is also connected to an extremal dipole ring of @xcite .
the dipole ring on this branch is such that in the infinite radius limit , the boost along the string _ preserves _ supersymmetry .
black ring solutions on this branch are also not known in full generality .
a five parameter family of black rings is conjectured to exist in minimal supergravity @xcite . for ranges of parameters
for which this family would admit the infinite radius limit , the corresponding string solution would be contained in our general string .
after some manipulations , one can show that for our general black string the tension can be set to zero for two distinct values of the boost parameter when the remaining parameters are left arbitrary ( and finite ) .
the two distinct branches of pressureless strings are distinguished by the relative sign of @xmath48 and @xmath49 , or in five - dimensional language , by the property that the orientation of the momentum along the string is aligned or anti - aligned with the orientation of the magnetic charge @xmath50 .
we refer to these two branches as the _ aligned _ pressureless string and the _ anti - aligned _ pressureless string .
the former contains branch @xmath47 of the pressureless extremal strings while the latter contains branches @xmath43 and @xmath46 of the pressureless extremal strings .
see figure [ fig : rings ] .
let us examine the properties of these pressureless strings in a little more detail .
without loss of generality , we only consider @xmath51 and @xmath52 by choosing @xmath53 as explained in section [ sec : monster ] .
the aligned string has then @xmath54 and the anti - aligned string has @xmath55 .
after a careful numerical analysis , we find that the two branches obey the following bound m^_5 |p_z| , where @xmath56 is the mass of the neutral pressureless black string ( @xmath57 ) .
the fact that the ( absolute value of the ) linear momentum is bounded from below indicates that the two branches are always separated .
also , we note that for a fixed value of the mass and magnetic charge , the two branches of pressureless strings have a different electric charge which always obeys q_^ q_^ , m_5 , q _ , the electric charge simply vanish
. it would be interesting to have an intuitive understanding of this physical effect .
the presence of two distinct pressureless limits suggests that there are two classes of black rings each admitting an infinite radius limit .
black rings on different branches rotate in different directions relative to the orientation of the dipole charge .
see figure [ fig : rings ] .
the analysis of @xcite shows that there is no smooth near - horizon geometry that describes a supersymmetric black ring in ads
the near horizon geometry of a supersymmetric black ring in ads necessarily possesses a conical singularity . in @xcite
a more physical argument was presented for the absence of supersymmetric black rings in ads .
it was based on the observation that the ads potential requires pressure along the one - brane in order to achieve mechanical equilibrium when the one - brane is curved into a contractible circle in ads .
it was speculated that supersymmetry is presumably incompatible with this pressure , and so , in ads we should not expect supersymmetric black rings .
however , the three parameter supersymmetric black string of @xcite invalidates the argument made in @xcite .
the supersymmetric black string of @xcite can have non - zero pressure ( see ) , so one can indeed balance it in ads by curving it into a contractible circle .
however , upon curving a straight supersymmetric string , the configuration will , in general , not remain supersymmetric .
the extremal strings of @xcite and more generally our general string provide requisite candidate strings that might correspond to the infinite radius limit of some ( non-)extremal black rings in ads .
a detailed perturbative study following @xcite is left for the future .
we would like to thank stphane detournay , roberto emparan , axel kleinschmidt , gary horowitz and jan perz for discussions , and ella jamsin for collaboration at initial stages of the project .
this research is supported in part by the national science foundation under grant no .
phy05 - 51164 .
av is supported by iisn - belgium ( conventions 4.4511.06 and 4.4514.08 ) and by the belgian federal science policy office through the interuniversity attraction pole p6/11 .
the work of g.c . is supported in part by the us national science foundation under grant no .
phy05 - 55669 , and by funds from the university of california .
the work of s.db .
is funded by the european commission though the grants piof - ga-2008 - 220338 ( home institution : universit libre de bruxelles , service de physique thorique et mathmatique , campus de la plaine , b-1050 bruxelles , belgium ) .
the work of ss is supported by the natural sciences and engineering research council of canada .
ss would also like to thank the university of california , santa barbara for their gracious hospitality where this research was conducted .
s. d. mathur , `` the quantum structure of black holes , '' class .
* 23 * , r115 ( 2006 ) [ arxiv : hep - th/0510180 ] .
j. r. david , g. mandal and s. r. wadia , `` microscopic formulation of black holes in string theory , '' phys .
rept . * 369 * , 549 ( 2002 ) [ arxiv : hep - th/0203048 ] .
j. c. breckenridge , d. a. lowe , r. c. myers , a. w. peet , a. strominger and c. vafa , `` macroscopic and microscopic entropy of near - extremal spinning black holes , '' phys .
b * 381 * , 423 ( 1996 ) [ arxiv : hep - th/9603078 ] .
h. elvang , r. emparan , d. mateos and h. s. reall , `` a supersymmetric black ring , '' phys .
lett . * 93 * , 211302 ( 2004 ) [ arxiv : hep - th/0407065 ] .
a. bouchareb , g. clement , c. m. chen , d. v. galtsov , n. g. scherbluk and t. wolf , `` @xmath58 generating technique for minimal d=5 supergravity and black rings , '' phys . rev .
d * 76 * , 104032 ( 2007 ) [ erratum - ibid .
d * 78 * , 029901 ( 2008 ) ] [ arxiv:0708.2361 [ hep - th ] ] .
s. giusto and a. saxena , `` stationary axisymmetric solutions of five dimensional gravity , '' class .
* 24 * , 4269 ( 2007 ) [ arxiv:0705.4484 [ hep - th ] ] .
d. v. galtsov and n. g. scherbluk , `` generating technique for @xmath59 supergravity , '' phys .
d * 78 * , 064033 ( 2008 ) [ arxiv:0805.3924 [ hep - th ] ] .
p. figueras , e. jamsin , j. v. rocha and a. virmani , `` integrability of five dimensional minimal supergravity and charged rotating black holes , '' class . quant .
* 27 * , 135011 ( 2010 ) [ arxiv:0912.3199 [ hep - th ] ] .
g. bossard , h. nicolai and k. s. stelle , `` gravitational multi - nut solitons , komar masses and charges , '' gen .
grav . * 41 * , 1367 ( 2009 ) [ arxiv:0809.5218 [ hep - th ] ] .
t. harmark and n. a. obers , `` general definition of gravitational tension , '' jhep * 0405 * , 043 ( 2004 ) [ arxiv : hep - th/0403103 ] .
z. w. chong , m. cvetic , h. lu and c. n. pope , `` general non - extremal rotating black holes in minimal five - dimensional gauged supergravity , '' phys .
lett . *
95 * , 161301 ( 2005 ) [ arxiv : hep - th/0506029 ] .
j. p. gauntlett , j. b. gutowski , c. m. hull , s. pakis and h. s. reall , `` all supersymmetric solutions of minimal supergravity in five dimensions , '' class .
* 20 * , 4587 ( 2003 ) [ arxiv : hep - th/0209114 ] .
h. k. kunduri , j. lucietti and h. s. reall , `` do supersymmetric anti - de sitter black rings exist ?
, '' jhep * 0702 * , 026 ( 2007 ) [ arxiv : hep - th/0611351 ] .
h. k. kunduri and j. lucietti , `` near - horizon geometries of supersymmetric ads(5 ) black holes , '' jhep * 0712 * , 015 ( 2007 ) [ arxiv:0708.3695 [ hep - th ] ] . | using @xmath0 dualities we construct the most general black string solution of minimal five - dimensional ungauged supergravity .
the black string has five independent parameters , namely , the magnetic one - brane charge , smeared electric zero - brane charge , boost along the string direction , energy above the bps bound , and rotation in the transverse space .
in one extremal limit it reduces to the three parameter supersymmetric string of five - dimensional minimal supergravity ; in another extremal limit it reduces to the three parameter non - supersymmetric non - rotating extremal string of five - dimensional minimal supergravity .
it also admits an extremal limit when it has maximal rotation in the four - dimensional transverse space .
the decoupling limit of our general black string is a btz black hole times a two sphere .
the macroscopic entropy of the string is reproduced by the maldacena - strominger - witten cft in appropriate ranges of the parameters .
when the pressureless condition is imposed , our string describes the infinite radius limit of the most general class of black rings of minimal supergravity .
we discuss implications our solution has for extremal and non - extremal black rings of minimal supergravity .
` ulb - th/10 - 23 ` * * geoffrey compre@xmath1 , sophie de buyl@xmath1 , sean stotyn@xmath2 , and amitabh virmani@xmath3 _ _ @xmath1 university of california at santa barbara + ca93106 santa barbara , united states + [email protected]@xmath4 [email protected] @xmath5 + @xmath5 + @xmath6 department of physics and astronomy , university of waterloo , + waterloo , ontario n2l 3g1 , canada + [email protected] + @xmath5 + @xmath7 physique thorique et mathmatique , + universit libre de bruxelles and international solvay institutes + campus plaine c.p .
231 , b-1050 bruxelles , belgium + [email protected] |
the interaction between an atom and an ion is caused by an induced atomic dipole due to the electric field of the ion . at large distances
it is given by : @xmath0 with @xmath1 .
here , @xmath2 is the charge of the ion and @xmath3 is the static polarisability of the atom .
it is useful to introduce the length scale @xmath4 and the energy scale @xmath5 $ ] that characterize the atom - ion potential , with @xmath6 the reduced mass . for the atom - ion combinations studied in this work we have that @xmath7 nm and @xmath8 hz for the atom / ion pair @xmath9rb/@xmath10yb@xmath11 @xcite , and @xmath12 nm and @xmath13 khz for the atom / ion pair @xmath14li/@xmath10yb@xmath11 @xcite . for @xmath15 , eq .
( [ eqpot ] ) does not describe the potential anymore as it becomes strongly repulsive .
the exact form of the potential in this regime is generally not known well enough to solve the scattering dynamics , but as it has been shown in ref .
@xcite , qdt can be employed to parametrize the potential at short - range ( see ref .
@xcite for more details ) .
although the potential ( [ eqpot ] ) is clearly spherically symmetric , we will for now limit ourselves to the one - dimensional case and we will denote the position of the ion ( atom ) as @xmath16 ( @xmath17 ) .
we note that in 1d it turns out that the atom - ion interaction has the same mathematical expression as in 3d , that is @xmath18 @xcite .
we discuss the 3d scenario in the appendix and compare it to the 1d results . as an illustration of the atomic dynamics in a double - well system in the presence of an ion
, we will first solve the system assuming that the ion is pinned to the center of its trap at @xmath19 .
hence , the hamiltonian of the atom is : h_a=+v_dw(z_a ) - , with @xmath20 the momentum of the atom , @xmath21 the atomic mass and @xmath22 the double - well potential . a convenient choice for @xmath22 is given by [ eq : vdw ] v_dw(z_a)=(z_a^2-d^2)^2 .
this potential has minima at @xmath23 with local trapping frequencies @xmath24 and inter - well barrier @xmath25 . for the sake of simplicity and without loss of generality , we fix the local trapping frequency @xmath26 for each inter - well distance @xmath27 by setting @xmath28 .
our goal is to find a set of basis functions to expand the solution of @xmath29 onto , as this procedure is more efficient than solving the schrdinger equation for each @xmath30 separately . to this end , we write the hamiltonian as @xmath31 with : @xmath32 note that the term @xmath33 has been added to ( [ eq_ha0 ] ) and subtracted again in eq . ( [ eq_ha1 ] ) .
this allows for finding a set of discrete basis states instead of the continuum that arises without this term for @xmath34 .
this is convenient as the final solutions of @xmath29 will form a discrete set as well @xcite . to solve the schrdinger equation for @xmath35 by means of qdt
, we note that as @xmath36 the energy is dominated by the term @xmath37 .
therefore , we can neglect the other energies and solve the schrdinger equation analytically .
we obtain even and odd solutions given by @xcite : @xmath38 where @xmath39 and @xmath40 are the even and odd short - range phases , respectively .
since the short - range phases are not generally known experimentally and can not be reliably obtained from ab - initio calculations , we choose a number of realistic values here , corresponding to scattering lengths in the range @xmath41 to @xmath42 .
we also note that in a quasi 1d setup the values of the phases can be tuned by changing the confinement in the two remaining dimensions @xcite . in order to solve the schrdinger equation for the hamiltonian ( [ eq_ha0 ] ) , we use eqs .
( [ eq_asymp_a],[eq_asymp_b ] ) as a boundary condition at some small value @xmath43 such that @xmath44 , with @xmath45 being the largest energy considered in the problem .
a renormalized numerov method @xcite gives the solutions @xmath46 with energies @xmath47 and quantum number @xmath48 . to solve the dynamics in the double - well we now diagonalize the hamiltonian with matrix elements : @xmath49 with @xmath50 for a range of values @xmath30 .
( color online ) upper panel : eigenenergy spectrum as a function of the inter - well separation @xmath30 for @xmath9rb and @xmath10yb@xmath11 assuming @xmath51 , @xmath52 ( black solid lines ) and @xmath53 , @xmath54 ( blue dashed lines ) . for large well separations
, the spectrum resembles that of two independent harmonic oscillators of the same trap frequency @xmath551.8 khz .
as the wells approach , each level splits in two owing to the energy difference of even and odd wavefunctions . in the right corner a molecular state
can also be seen .
the occuring tunnelling rate is given by the energy difference and is plotted in the lower panel .
, width=302 ] now , as in our previous study @xcite , we consider @xmath9rb and @xmath10yb@xmath11 as an example system and we assume that no spin - changing collisions can occur such that our single channel description is accurate . setting @xmath55
1.8 khz we obtain the spectrum shown in fig .
[ fig : atomonlyspec ] . for large inter - well separations
@xmath30 the eigenenergies resemble the equidistant level structure of the harmonic oscillator as the wells are uncoupled .
while decreasing the inter - well distance , the energy levels split in two as there is an energy difference between states of even and odd symmetry . for a single atom and small energy splitting
, this energy difference corresponds to the inter - well tunnelling rate in a two - mode picture @xcite . since the barrier height @xmath25 decreases with distance , the higher excited atomic states exhibit level splitting at larger separation @xmath30 than low energy states .
the eigenenergies that asymptotically connect to the harmonic oscillator ground states correspond to superpositions of approximately localized wavepackets : @xmath56 in the left ( l ) and right ( r ) well .
these states are labeled with the subscripts @xmath57 and @xmath2 and have energies @xmath58 with @xmath59 . at @xmath60 900 nm
we find that the ground state degeneracy is lifted corresponding to a tunnelling rate of @xmath61 56 hz when @xmath51 , @xmath52 ( black lines ) and @xmath62 202 hz when @xmath53 , @xmath54 ( dashed blue lines ) .
since the short - range phases depend on the relative spin orientations of the atom and ion , the above calculation demonstrates a similar ion spin - dependence as in ref .
finally , we note that the spectrum was obtained by taking 106 basis states into account . in appendix [ app_3d_m ] we perform a 3d analysis of the double - well for a static ion that we compare to the 1d model of the present section .
we show that the double - well problem in 1d has a strict analogy to the 3d scenario , and therefore it provides a satisfactory physical picture of the system . given this ,
we extend our analysis to the case of a moving ion .
a 3d study of such a system would be indeed rather difficult to treat numerically and nonetheless it would not provide further insight into the problem .
now we will see how the picture is altered when we allow the ion to move . as outlined above
, we will restrict our attention to the 1d scenario . in this case the hamiltonian is given by : [ h_dw_moving ] h=++m_i_i^2z_i^2+v_dw(z_a)- . here
, @xmath63 denotes the ion trap frequency , @xmath64 the mass of the ion , and @xmath65 its momentum . to find the eigenstates and energies of this hamiltonian
, we write the hamiltonian in terms of relative and center - of - mass coordinates @xmath66 , @xmath67 with @xmath68 the total mass , as @xmath69 : @xmath70 equation ( [ eq_h0r ] ) denotes the hamiltonian of a harmonic oscillator with trap frequency @xmath71 .
it has eigenstates @xmath72 and energies @xmath73 .
equation ( [ eq_h0r ] ) is similar to equation ( [ eq_ha0 ] ) except for some pre - factors , in particular @xmath74 .
as explained in the previous section , we have determined the eigenfunctions @xmath75 and eigenenergies @xmath76 of such hamiltonian by means of qdt . to find the eigenstates and eigenenergies to the full hamiltonian we expand onto the basis @xmath77 .
we denote the solutions for each inter - well separation @xmath30 as @xmath78 with energies @xmath79 and coefficients @xmath80 , and where an overall quantum number @xmath81 labels each solution . as an example of the resulting spectrum
we show the situation for @xmath9rb and @xmath10yb@xmath11 assuming @xmath51 , @xmath52 , @xmath82khz and @xmath83khz in fig . [
fig : ionmovingspecwithandwithoutia ] .
this spectrum was obtained by expanding the solutions onto 6417 basis states ( 93 states in the relative coordinate and 69 states in the center - of - mass coordinate ) . in comparison to fig .
[ fig : atomonlyspec ] we see that more energy levels appear , both in the form of molecular states and trap states .
when the wells are far apart , the states reduce to the harmonic oscillator states for the atom and ion .
we can identify the asymptotic atom - ion fock states @xmath84 in fig .
[ fig : ionmovingspecwithandwithoutia ] for large @xmath30 . for intermediate @xmath30 ,
the states are perturbed by the atom - ion interaction .
focusing on the states that connect to the harmonic oscillator ground states @xmath85 , we see the same behaviour as for the case where the ion was static : the initial degeneracy for large inter - well separations is lifted as the wells come closer and a tunnelling rate can be identified .
molecular states cross the ground states at a few points , but the avoided crossings are very small .
therefore , the static ion approximation was justified when considering the ground states .
the state connecting to @xmath86 shows a similar behaviour , but more and larger avoided crossings appear .
close to these crossings , a simple two - mode picture breaks down and the tunnelling rate @xmath87 loses its meaning . in an experiment ,
the inter - well separation @xmath30 would typically be reduced dynamically to initiate tunnelling .
the resulting tunneling rate then depends on whether the crossing is traversed diabatically or not . for most values of @xmath30 ,
the energy splitting is different than for the case when the ion is in the ground state . for @xmath88
nm , for instance , we get a tunnelling rate of @xmath89 hz when the ion is in the ground state , but only @xmath90 hz when the ion is in the first fock state .
therefore , the atomic tunnelling rate depends on the motional state of the ion . as a second example , we also plot the case for @xmath14li and @xmath10yb@xmath11 assuming @xmath51 , @xmath52 , @xmath91 khz and @xmath92 khz ( see fig .
[ fig : ionmovingspecybli ] ) .
we see that the tunnelling occurs for larger inter - well separations for these species as the atomic wavepacket is larger because of its lower mass .
furthermore , the tunnelling rate shows less dependency on the motional state of the ion as long as there are no avoided crossings nearby . this is a consequence of the large mass ratio between the atom and ion that separates atomic and ionic dynamics in a similar fashion as in the born - oppenheimer approximation .
we note that in this case 2700 states were taken into account . in fig .
[ fig : wavs2d_q2p56_rb_yb_-pi4_pi4_n0m0 ] we plot a few probability distributions for the eigenstates corresponding to the red dots in fig . [
fig : ionmovingspecwithandwithoutia ] .
we see that for @xmath93 , the atom - ion interaction dominates , but for larger atom - ion distances the probability distributions resemble those of the harmonic oscillator in each well and the presence of the ion only slightly perturbs the double - well system ( color online ) upper panel : spectrum of a double - well system where the ion is allowed to move for @xmath9rb and @xmath10yb@xmath11 with @xmath82khz , @xmath83khz , @xmath51 , and @xmath52 . on the right
, we give some of the asymptotic quantum numbers @xmath94 for large well separation .
the red dots correspond to the wavefunctions plotted in fig .
[ fig : wavs2d_q2p56_rb_yb_-pi4_pi4_n0m0 ] . in the lower panel ,
the level splitting is plotted for the case where the ion is in the ground state ( thick black line ) and for the case where it is in the first excited state ( dashed lines ) .
when avoided crossings occur , the energy level separation becomes dependent on whether the crossing is traversed diabatically or adiabatically .
the smallest @xmath87 is plotted using a thicker dashed blue line .
close to the avoided crossings , the two mode description is no longer accurate and @xmath87 loses its interpretation as the tunnelling rate.,width=321 ] ( color online ) upper panel : spectrum of a double - well system where the ion is allowed to move for @xmath14li and @xmath10yb@xmath11 with @xmath82khz , @xmath83khz , @xmath51 , and @xmath52 , i.e. the same as in fig .
[ fig : ionmovingspecwithandwithoutia ] . note that tunnelling occurs for larger distances than for @xmath9rb as the @xmath95li wavepacket is larger for the same trapping frequency . on the right , we give the asymptotic quantum numbers @xmath94 for large well separation .
the lower panel shows the energy level separation @xmath87 for the ion in the ground state ( black solids line ) and the first excited state ( dashed blue ) .
the avoided crossings are also indicated by the thin grey lines .
clearly , in this case the tunnelling rate is very similar for the first excited and ground states when no avoided crossings are near .
, width=321 ] ( color online ) wavefunctions @xmath96 of the ion - atom states for an inter - well separation of @xmath97=765 nm for @xmath9rb and @xmath10yb@xmath11 with @xmath51 , @xmath52 , @xmath98khz , and @xmath99khz , corresponding to the red dots in fig .
[ fig : ionmovingspecwithandwithoutia ] .
we used units of @xmath42 on the axes as these are the natural units for the atom - ion system .
the states asymptotically correspond to the fock states @xmath100 , @xmath101 , @xmath86 and @xmath102 for large well separation ( with the atom occupying both wells ) . the top wavefunction is the double - well ground state @xmath103 .
the corresponding quantum numbers can still be recognised by counting the nodes in the direction of the ionic and atomic coordinate . in the vicinity of the line
@xmath104 the atom - ion interaction potential dominates and the wavefunctions take similar forms as in eqs .
( [ eq_asymp_a],[eq_asymp_b ] ) .
the fast oscillations around this region are not completely resolved in the density plot .
, width=283 ] rb and @xmath10yb@xmath11 with @xmath51 , @xmath52 , @xmath98khz and @xmath99khz .
, width=302 ] an interesting application using the dependence of the tunnelling rate on the motional state of the ion is the generation of entanglement between the atom and the ion . in ref .
@xcite the state dependent tunnelling was proposed via the state - dependence of the short - range phase .
alternatively , we can create non - classical states of ion motion , such as fock states , to engineer motional state - dependent tunnelling .
the fock state can in turn be entangled with the spin of the ion .
such quantum states are commonly engineered and analyzed in ion trap experiments @xcite . from the spectrum displayed in fig [ fig : ionmovingspecwithandwithoutia ] , we see that a good strategy to engineer atom - ion entanglement consists of the following steps : 1 .
we first prepare the ion in an equal superposition of the ground and first excited state of its trap , while its internal state is prepared in a specific hyperfine level .
2 . subsequently , we prepare the atom , for instance , in the ground state of the left well and far away from the ion such that the tunnelling is negligible .
then we reduce the inter - well separation @xmath30 dynamically until tunnelling occurs .
4 . afterwards
, we wait until the atom has tunnelled to one side depending on the ion fock state , that is , to the right well if the ion fock state is @xmath105 and back to the left well if the ion fock state is @xmath106 ( see fig .
[ fig : ionmovingspecwithandwithoutia ] first and third red dots , from below ) . for the ion in the ground state , @xmath107 , we see that the sequence is very similar to the one presented in ref .
@xcite ( see fig . [
fig : ionmovingspecwithandwithoutia ] ) . for higher fock states ,
however , we see an increasing amount of avoided crossings .
to pass these crossings diabatically during our sequence , an accurate control of the inter - well distance @xmath108 is required .
we see that for @xmath109 the number of crossings is still quite limited and we take this as an example .
we have performed numerical simulations of the dynamics and we have optimised the process , by means of the chopped random - basis algorithm @xcite , in order to maximise the entanglement . to this aim , we have defined the following overlaps : @xmath110 here , @xmath111 is the wave function of the atom in the ground state of the right well and the ion in the ground state of its harmonic trap , whereas @xmath112 is the wave function of the atom in the ground state of the left well and the ion in the first excited state of its harmonic trap .
the time evolved states are obtained from @xmath113 and @xmath114 , where @xmath115 is the time evolution operator generated by the hamiltonian ( [ h_dw_moving ] ) . in the optimization procedure we minimised the overlap infidelity , @xmath116 , at the final time @xmath117 , that is , the time needed to perform the sequence 3.-4 . outlined above , by optimally controlling the inter - well separation @xmath108 . in fig .
[ fig : entanglement ] we show the overlaps @xmath118 for the optimal inter - well separation @xmath108 .
we obtained @xmath119 and @xmath120 with @xmath121 ms for the atom - ion pair @xmath9rb/@xmath10yb@xmath11 .
this result shows that we can produce atom - ion entanglement in a very short time by means of the ionic motional state .
we note , however , that the overlaps are not perfect .
this might be achieved by controlling , for instance , the atom - ion interaction via feshbach resonances , or by controlling additionally the height of the barrier .
our goal here , however , is to show that such entanglement generation is in principle possible
. a more detailed analysis by means of optimal control theory would require perfect knowledge of the trapping potentials , and not just the analytically convenient form given by eq .
( [ eq : vdw ] ) .
( color online ) optimisation of the entanglement generation protocol as discussed in the main text .
upper panel : overlaps @xmath118 for the optimised dynamics .
lower panel : optimal inter - well separation .
the optimisation has been performed for the atom - ion pair @xmath9rb/@xmath10yb@xmath11 with @xmath51 and @xmath52 . ] up until now , we have assumed that both the atom and the ion are prepared in a pure state by ground state cooling ( for instance via resolved sideband cooling ) followed by coherent state manipulation .
to see what the effect of imperfect ground state cooling of the ion would be , we have plotted in fig .
[ fig : spec_rbyb_gam5p5_-pi4_pi4_alpha10_n_all ] the spectra for the first 5 states that asymptotically correspond to the first 5 fock states of the ion .
when the ion is in a thermal state , each of these fock states contributes to the tunnelling rate in a incoherent manner .
we see that for @xmath122 the energy levels display degeneracy in @xmath123 for large @xmath30 , in analogy to the situation for the ground states discussed above . as the inter - well separation
is decreased these degeneracies are lifted and tunnelling occurs . because of the coupling between the atomic tunnelling and the motion of the ion , each ionic state corresponds to a different tunnelling rate .
unless only few fock states are occupied in the thermal state or the tunnelling rates are very similar for all fock states , the coherence in the atomic state will be destroyed after a while . for the states of higher energy @xmath124 we see that many avoided crossings start appearing that can not be ignored anymore for small @xmath30 .
in this situation , the simple two mode picture can not be employed anymore . as an example , we have focused our attention to the tunnelling only , in particular when the ion is ( ideally ) prepared in the ground state . to this end
, we performed first an optimisation of this dynamics .
the corresponding result is shown in fig .
[ fig : tunnelling ] . in this case
we were able to achieve an overlap fidelity of about @xmath125 in @xmath121 ms . given this result , we have investigated the impact of finite temperature on the tunnelling dynamics . to begin with ,
we have defined the initial density matrix of the ion as @xmath126 where @xmath127 are the harmonic oscillator eigenstates of the ionic harmonic trap , @xmath128 are the occupation probabilities which are calculated by assuming a thermal distribution corresponding to temperature @xmath129 in the canonical ensemble .
the overlap fidelity at time @xmath130 for a given temperature is given by @xmath131 here , @xmath132 is the time evolved state assuming that the initial state is the ground state of the left well for the atom and the @xmath133-th state of the harmonic trap for the ion .
it is interesting to see values of the fidelity for temperatures up to @xmath134 ( @xmath135 is the boltzmann constant ) .
let us define @xmath136 and neglect terms of @xmath137 in eq .
( [ eq : ft ] ) , since for higher energies the motional states of the atom and the ion are not anymore separable . such temperature dependence of the fidelity is displayed in fig .
[ fig : temperature ] , for which we used the optimal inter - well separation shown in the lower panel of fig .
[ fig : tunnelling ] .
as it is shown , the fidelity drops rather quickly , since the overlaps at time @xmath117 for the ionic states @xmath138 are significantly reduced .
this shows that the tunnelling rates are different for those fock states , and therefore the coherence in the atomic state is destroyed rapidly .
a strategy to reduce the impact of finite temperature could be to perform an optimisation in which the optimal inter - well separation @xmath108 is engineered in such a way that @xmath139 is maximised within a given temperature range .
this decoherent dynamics will be studied in the future more systematically via a quantum open system approach .
( color online ) optimisation of the tunnelling dynamics when the ion is prepared in the ground state .
upper panel : overlap @xmath140 for the optimised dynamics .
lower panel : optimal inter - well separation .
the optimisation has been performed for the atom - ion pair @xmath9rb/@xmath10yb@xmath11 with @xmath51 and @xmath52 .
the final overlap fidelity is @xmath141 . ]
( color online ) fidelity @xmath139 of the tunnelling dynamics for the atom - ion pair @xmath9rb/@xmath10yb@xmath11 , whose overlap fidelity when the ion is prepared in the ground state of the trap and the corresponding optimal inter - well separation are displayed in fig .
[ fig : tunnelling ] . ]
so far , we have assumed that the ion is trapped in a time - independent harmonic trap . in most atom - ion experiments ,
however , the ions are trapped in a paul trap that is based on an oscillating electric field .
this causes a rapid oscillating motion of the ion called micromotion .
the hamiltonian describing this situation in 1d is given by : @xmath142.\ ] ] here , the parameters @xmath143 and @xmath144 depend on the trapping potential , the ion mass and its charge .
with properly chosen @xmath143 and @xmath144 , an effective time - independent trapping potential can be derived via the so - called secular approximation . in this approximation
the ion motion reduces to a combination of the slow ( secular ) motion , that is , of a particle in a time independent trap , and a rapid micromotion of small amplitude . in most experiments
the following inequalities hold : @xmath145 , where @xmath144 lies typically in the range 0.1 - 0.4 . in order to evaluate the impact of the time - dependent ion trap on the double - well system , we replace the ion trap term ( i.e. , the kinetic and trapping potential energies ) in eq . ( [ h_dw_moving ] ) by the time dependent version of eq .
( [ h_ion ] ) .
et al . _
@xcite have analysed a situation in which the atom was trapped in a harmonic trap . here
we will use the same approach , but for a double - well potential .
we perform the transformation of cook and shankland @xcite to obtain a hamiltonian that is comprised of a time - independent and a time - dependent term that we label with @xmath146 ( micromotion ) @xmath147 .
note that @xmath148 is the hamiltonian of the double well without micromotion , which is defined in eqs .
( [ eq_h0r],[eq_h0r],[eq_h1 ] ) .
more details on this procedure can be found in the appendix [ app_mm ] .
the micromotion hamiltonian in the center - of - mass and relative coordinates is given by : @xmath149 here @xmath150^{-1/2}$ ] and @xmath151 denotes the anti - commutator .
the relative and center - of - mass momenta are denoted by @xmath152 and @xmath153 , respectively .
following again ref .
@xcite , we use floquet theory to obtain the energies and eigenstates in terms of the unperturbed eigenstates of the hamiltonian @xmath148 . it turns out that we have to diagonalise the hamiltonian h_f = h^(d)+h_mm(t)-i .
we use the unperturbed floquet eigenstates @xmath154 of @xmath155 as our basis with floquet energies @xmath156 . here
the integer @xmath157 denotes the class of the floquet state , whereas the quantum number @xmath81 denotes the solution to the double - well problem without micromotion .
then , we introduce the generalised matrix elements : [ eq : genmatrxel ] u^*_jl|h_mm(t)|u_jl = _ 0^t dt u^*_jl|h_mm(t)|u_jl , where now @xmath117 indicates the period of the micromotion . to gain further insight into the micromotion effect we write the matrix elements as follows @xcite ( see appendix [ app_mm ] ) : @xmath158e^{i(j - j')\omega t}|\psi^{(d)}_{l}\rangle,\nonumber\\ v_1 & = & -m_i g^2\omega_i^2\left(r^2+\frac{m_a^2}{m^2}r^2+\frac{2m_a}{m}rr\right)\label{eq_v1},\\ v_2 & = & -\frac{i g \omega_i m_i}{\hbar}\left(e_{l'}-e_{l}\right)\left(r^2+\frac{m_a^2}{m^2}r^2+\frac{2m_a}{m}rr\right).\label{eq_v2}\end{aligned}\ ] ] the larger these matrix elements are , the less we expect the secular approximation to hold .
additionally , the coupling between two states may become resonant for states belonging to different floquet classes when @xmath159 .
we note that the selection rules for the coupling between different floquet classes are given by @xmath160 for @xmath161 , and @xmath162 for @xmath163 [ see also eq .
( [ eq : genmatrxel ] ) ] .
when the micromotion hamiltonian can not be considered a perturbation , the problem has to be solved by taking a large amount of floquet classes into account .
however , for large inter - well separation , we do not expect the micromotion to play a significant role as the atom is trapped too far away from the ion to sense the small amplitude micromotion of the ion . as @xmath30 is decreased
, we expect the micromotion term to become increasingly important . to quantify this , we will study how strongly the micromotion couples the states of interest @xmath164 to states with a different energy as @xmath30 is decreased .
we limit the discussion to @xmath165 as it can be shown that the effects are very similar for @xmath166 .
the energy difference prefactor @xmath167 causes @xmath161 to be the main perturbing term @xcite and we focus on it from now on .
let us introduce the following notation for the ( absolute value of ) these matrix elements : @xmath168 in fig .
[ fig : matelems ] a)-c ) the resulting matrix elements for the case discussed in fig .
[ fig : ionmovingspecwithandwithoutia ] are shown ( i.e. , for the atom - ion pair @xmath9rb/@xmath10yb@xmath11 ) . here
, the matrix elements are plotted for @xmath169 and 820 nm and the detuning of the coupling state @xmath170 .
clearly , the couplings are largest to states that are nearby in energy . for @xmath171 ,
the couplings quickly fall off to zero as the coupling state is further separated in energy .
the terms involving @xmath172 and @xmath173 , however , have significant couplings with states that are far separated in energy .
this behaviour is a direct consequence of the non - linear interaction between the atom and the ion . in the center - of - mass
coordinate the basis functions are fock states and the matrix elements involving only @xmath174 have well - defined selection rules : @xmath175 for @xmath176 and @xmath177 for @xmath178 .
since the atomic and ionic ground state are only slightly perturbed by the atom - ion interaction , as it is clear from section [ sec_moving ] , few fock states are involved and due to the selection rules , only coupling to nearby states occur . because there are also selection rules on @xmath157 , the coupling states need to be of a different floquet class .
these states are unlikely to become resonant since @xmath179 .
this situation is comparable to the effect that micromotion has on a single trapped ion . for the atom - ion scattering states in the relative
coordinate , however , there are no selection rules owing to the non - linear atom - ion interaction . for instance , @xmath180 in general for any @xmath48 and @xmath181 that have the same symmetry .
thus , the states of interest couple to highly excited states belonging to different floquet classes . for the case studied here , @xmath182 70.5 khz , which corresponds to @xmath183 . for inter - well distances of about 820 nm , the couplings to states that are @xmath184 70 khz separated in energy - and can therefore become resonant - reach up to 10% of the atomic trapping frequency
. such large couplings cause significant deviations from the secular solution if resonances occur .
this situation is indeed quite similar to the case studied in ref .
hence , we expect a large number of energy levels crossing the ground states with increasing strength as @xmath30 is reduced .
this will render the state preparation more difficult , when considering schemes where @xmath30 is slowly reduced to initiate atomic tunnelling .
only superb experimental control over @xmath108 will allow the diabatic transfer over avoided crossings such that atomic excitations within the wells are prevented .
it is interesting to note that the prefactors to the terms containing @xmath172 and @xmath173 in equations ( [ eq_v1 ] ) and ( [ eq_v2 ] ) , that will cause the largest effects , are @xmath185 and @xmath186 , respectively .
this suggests that adverse micromotion effects may be reduced by choosing @xmath187 , this observation is in line with the classical study of cetina _
_ @xcite concerning the limits of atom - ion sympathetic cooling . as a comparison for the case of the double - well , we plot also the matrix elements for the case of @xmath95li and @xmath10yb@xmath11 [ see fig . [
fig : matelems ] d)-f ) ] .
we see that the matrix elements are indeed smaller for this case and we reach a tunnelling rate of @xmath188 hz with micromotion induced couplings on the percent level of the atomic trap frequency or @xmath184 20 hz .
finally , we have numerically solved the full double - well problem including micromotion . the required total hilbert space dimensions scale as @xmath189 - that is the number of center - of - mass states to be taken into account , times the number of relative states times the number of floquet classes .
clearly , the calculation becomes very hard already at small @xmath30 .
for instance , calculating the eigenenergies for the case presented in fig .
[ fig : ionmovingspecwithandwithoutia ] taking only the classes @xmath190 into account , would require a hilbert space of dimension @xmath191 .
for the case of @xmath95li and @xmath10yb@xmath11 of fig .
[ fig : ionmovingspecybli ] the situation is better and we diagonalised the hamiltonian for 500 values of @xmath30 taking @xmath190 , such that 13500 states are used to form the basis .
the result is shown in fig .
[ fig : mmliyb ] a ) .
we can see that although many more energy levels are present in the spectrum , the coupling to the states @xmath192 is indeed very small , so that the micromotion should not pose a problem for the parameters considered here . as a second example
, we calculate the spectrum around the ground states for a higher trap drive frequency .
to reduce the numerical complexity we use @xmath95li and @xmath10yb@xmath11 and set all energy scales to similar values , i.e. @xmath193 khz , @xmath194 khz for @xmath183 and @xmath195 such that @xmath196 khz , we obtain the eigenenergy spectrum shown in fig .
[ fig : mmliyb ] . for this calculation we took @xmath197 , @xmath198 , and @xmath199 .
thus , the hilbert space comprised of 8640 basis states . as is illustrated in fig . [
fig : mmliyb ] , that also for this case , we see that the avoided crossings remain relatively small , and therefore enabling us to apply the secular approximation for the double - well system at large enough separations @xmath30 . since the avoided crossings remain small , we do not expect that taking more floquet classes into account will significantly alter the results for the values of @xmath30 plotted . on the other hand , for smaller @xmath30
, the ground state presents many avoided crossings , and therefore more floquet classes are required to reach convergence .
\a ) a ) spectrum around the states @xmath192 for @xmath95li and @xmath10yb@xmath11 with @xmath200 khz , @xmath201 khz for @xmath183 and @xmath195 such that @xmath202 khz . for the calculation we took @xmath197 , @xmath51 , and @xmath52 . to construct the basis 13500 states were used .
the hamiltonian was diagonalized for 500 different values of @xmath30 .
it can be seen that many energy crossings appear but the avoided crossings remain very small .
b ) spectrum for @xmath95li and @xmath10yb@xmath11 with @xmath193 khz , @xmath194 khz for @xmath183 and @xmath195 such that @xmath203 137 khz . for the calculation we took @xmath197 , @xmath198 , and @xmath199 . to construct the basis 8640 states were used .
the hamiltonian was diagonalized for 600 different values of @xmath30 .
it can be seen that many energy crossings appear and that the avoided crossings become bigger as @xmath30 decreases .
, title="fig:",width=302 ] b ) a ) spectrum around the states @xmath192 for @xmath95li and @xmath10yb@xmath11 with @xmath200 khz , @xmath201 khz for @xmath183 and @xmath195 such that @xmath202 khz . for the calculation we took @xmath197 , @xmath51 , and @xmath52 . to construct the basis 13500 states were used .
the hamiltonian was diagonalized for 500 different values of @xmath30 .
it can be seen that many energy crossings appear but the avoided crossings remain very small .
b ) spectrum for @xmath95li and @xmath10yb@xmath11 with @xmath193 khz , @xmath194 khz for @xmath183 and @xmath195 such that @xmath203 137 khz . for the calculation we took @xmath197 , @xmath198 , and @xmath199 . to construct the basis 8640
states were used .
the hamiltonian was diagonalized for 600 different values of @xmath30 .
it can be seen that many energy crossings appear and that the avoided crossings become bigger as @xmath30 decreases .
, title="fig:",width=302 ]
atomic double - well potentials can be created by using optical tweezers or standing wave laser fields @xcite .
such fields are sufficiently strong to also trap an ion , as it has first been shown in a recent proof - of - principle experiment @xcite .
an all optical or hybrid optical / paul trap would solve the issues related to micromotion . in reference @xcite , for instance
, a single ion was trapped in a standing wave laser field .
it is feasible to alter this setup to allow both a harmonic potential for the ions and a multi - well potential for the atoms by overlapping trapping beams of suitable frequencies .
hence , the system studied in this paper may be realised by using optical trapping fields , or by means of a hybrid approach in which a paul trap is combined with an optical one .
atomic double - well systems have been created in atomic microtraps , or atom chips @xcite , where atoms are magnetically trapped by means of integrated current carrying wires @xcite .
planar ion traps have also become available in recent years @xcite , whose development was mainly driven by the prospect of developing a scalable quantum computer @xcite .
it seems feasible to combine the technologies of atom chips and planar ion traps to implement the proposed setup .
recently , for instance , magnetic field gradients were used in planar ion traps for spin detection or to design spin - spin interactions @xcite .
those traps combine electrodes on the surface of a chip that form the ionic trap with current carrying wires to produce magnetic gradient fields .
gradients ranging from 1 - 20
t / m at @xmath184100 @xmath6 m above the surface can be reached .
slight modifications to such a setup would lead to magnetic field patterns that can trap atoms .
for instance , we have calculated that by using a layered design , in which a planar ion trap is mounted on top of a current carrying structure , results in magnetic gradients of about 7 t / m . for atoms trapped 100 @xmath6 m above the surface
, this would correspond to trapping frequencies in the khz range @xcite with appropriately aligned bias fields .
such a setup could be combined with radio - frequency fields as well , leading to adiabatic dressed double - well potentials , in analogy to the works of refs .
in such a setup , the inter - well separation could be dynamically tuned by changing the frequency of the dressing field .
we have theoretically studied the dynamics of an atomic double - well system in the presence of a single trapped ion . we have found , under the assumption that the ion is not moving , that a one - dimensional calculation provides the same physical picture as a three - dimensional one .
the spin of the ion can control the tunnelling rate via the state dependent short - range phase , as also discussed in reference @xcite .
when the ion is allowed to move , we find that the atomic tunnelling rate between the wells also couples to the ion motion . in this way
, the tunnelling can be controlled by the motion of the ion . as the motional state of trapped ions
is routinely engineered and read - out in state - of - the - art experiments @xcite by coupling it to its internal ( spin ) state , this may allow engineering atom - ion states in which the atomic position is entangled with the ion motion or spin . we have analyzed a scheme in which the inter - well distance is dynamically reduced to allow the atom to tunnel depending on the motional state of the ion .
imperfect ion ground state cooling will result in reduced tunnelling contrast .
since ion heating may occur on the timescale of the tunnelling dynamics in experiments where the ion is trapped close to the electrodes , we plan to analyze the effect of ion heating and cooling on the tunnelling dynamics in the future .
the coupling between atomic tunnelling and ion motion can also be seen as a unit - cell for a larger atom - ion quantum simulator , in which the atomic dynamics is coupled to phonons in an ion crystal @xcite .
we have also analyzed the effect of micromotion on the energy spectrum of the double - well system .
the micromotion causes many extra avoided crossings in the spectrum as coupling to states belonging to different floquet classes becomes possible for small inter - well separation @xmath30 .
the exact strength of these crossings depends on the trap parameters , but we conclude that it is a good idea to choose an ion - atom combination with a large mass ratio .
the avoided crossings will cause trouble in state preparation as the wells have to be brought together diabatically with respect to the crossings , but without exciting the atoms to higher trap states .
additionally , the two - mode approximation may break down in a many - body scenario when including the micromotion , complicating its theoretical description . in linear paul traps ,
the dynamical electric field is only used to confine the ions in two of the three directions and it may be advisable to have the double - well separation in the third direction where the ion is confined by static fields . nonetheless , due to the spherical symmetry of the atom - ion interaction potential , there will be some minimum distance where the secular approximation will fail @xcite .
we have given two possible routes towards experimental implementation of the double - well system .
either the double - well potential is created by using magnetic fields , or with optical fields .
both technologies are compatible with ion trapping , putting an experimental realization of the considered system within reach .
this work was supported by the erc ( grant 337638 ) ( r.g . and j.j .
) , by an internal grant of the university mainz ( project ` hybrid atom - ion micortrap ' ) ( r.g . and j.j . ) , and by the excellence cluster ` the hamburg centre for ultrafast imaging - structure , dynamics and control of matter at the atomic scale ' of the deutsche forschungsgemeinschaft ( a.n . ) .
the authors acknowledge useful discussions with andrew grier and krzysztof jachymski , and are grateful to ferdinand schmidt - kaler , tommaso calarco , and zbigniew idziaszek for previous work on a related project .
rg thanks ulrich poschinger for proof reading the manuscript .
the procedure for obtaining the 3d eigenenergies and states has been described in ref .
@xcite and for completeness we give more details here .
the 3d hamiltonian is given by @xmath204 , where @xmath205 here @xmath206 is the trapping frequency in the transverse direction .
we introduce spherical coordinates @xmath207 , where we kept the subscript @xmath143 for denoting the coordinates that belong to the atom .
the solutions to the radially symmetric equation ( [ eq_3d ] ) are of the form @xmath208 , where @xmath209 are spherical harmonics , and @xmath210 are the solutions to the radial schrdinger equation . since the potential does not depend on the azimuthal direction , the quantum number @xmath211 is conserved and for simplicity we set @xmath212 .
the radial schrdinger equation is given by : [ eqspherical ] e_nl^(0)^0_nl=(-++-)^0_nl.to numerically solve this equation , we make use of the renormalized numerov method @xcite with ^0_nl(r_a ) as a boundary condition , where @xmath213 .
the mixing angle @xmath214 is related to the 3d short - range phase as : @xmath215 .
we then expanded the solution of the full hamiltonian @xmath216 onto the solutions of eq .
( [ eq_3d ] ) , namely @xmath217 , where @xmath48 denotes the pair of quantum numbers @xmath218 .
thus , both the wavefunctions and the corresponding energies are obtained by diagonalising the hamiltonian @xmath216 , whose matrix elements are given by : @xmath219 with @xmath50 .
these matrix elements can be explicitly written as @xmath220 whose determination relies on the computation of the following clebsch - gordon coefficients : @xmath221}{(2l-3)(2l-1)(2l+3)(2l+5)},\nonumber\\ c^{(4)}_{l , l+2}&\!=\!&\frac{2(l+1)(l+2)[-3 + 2l(l+3)]\sqrt{5 + 4l(l+3)}}{(2l-1)(2l+1)(2l+3)(2l+5)(2l+7)},\nonumber\\ c^{(4)}_{l , l+4}&\!=\!&\frac{(l+1)(l+2)(l+3)(l+4)\sqrt{9 + 4l(l+5)}}{(2l+1)(2l+3)(2l+5)(2l+7)(2l+9)}.\nonumber\end{aligned}\ ] ] all other coefficients vanish due to selection rules on the quantum number @xmath81 . besides this , we note that the coefficient matrices are symmetric @xmath222 . in fig .
[ fig : atomonlyspec3d ] we show the spectrum for the parameters @xmath223 , @xmath224 khz , and @xmath200 khz , which is the same axial trapping frequency as used in fig .
[ fig : atomonlyspec ] .
we note that for such calculation we have obtained a set of 1838 eigenfunctions and energies including bound states as well as trap states @xcite . apart from the extra angular integration
, the whole procedure is very similar to the one used in the 1d calculation and the resulting energy spectrum looks indeed very similar .
there are , however , some differences .
first of all , since there are more degrees of freedom , more levels appear .
note also that not all levels are plotted since we limited our discussion to @xmath212 .
secondly , the exact way in which the energy levels split close to the ion is different .
this behavior , however , is related to the choice of the short - range phase . to determine the exact form experimental input
would be needed in the form of a scattering length .
three - dimensional eigenenergy spectrum as a function of inter - well separation @xmath30 for a @xmath9rb atom and a pinned @xmath10yb@xmath11 ion assuming a 3d short range phase of @xmath223 and @xmath200 khz , @xmath224 khz .
, width=302 ]
following refs .
@xcite we start by writing the ion wavefunction as : in order to evaluate the matrix elements of @xmath226 , that is , eqs .
( [ eq_v1],[eq_v2 ] ) , it is very useful to note that @xmath227|\psi^{(d)}_l\rangle/\hbar$ ] .
similarly , @xmath228|\psi^{(d)}_l\rangle/\hbar$ ] @xcite .
equivalent equations hold for the center - of - mass coordinate too . | we theoretically analyze the dynamics of an atomic double - well system with a single ion trapped in its center .
we find that the atomic tunnelling rate between the wells depends both on the spin of the ion via the short - range spin - dependent atom - ion scattering length and on its motional state with tunnelling rates reaching hundreds of hz .
a protocol is presented that could transport an atom from one well to the other depending on the motional ( fock ) state of the ion within a few ms .
this phonon - atom coupling is of interest for creating atom - ion entangled states and may form a building block in constructing a hybrid atom - ion quantum simulator .
we also analyze the effect of imperfect ground state cooling of the ion and the role of micromotion when the ion is trapped in a paul trap . due to the strong non - linearities in the atom - ion interaction ,
the micromotion can cause couplings to high energy atom - ion scattering states , preventing accurate state preparation and complicating the double - well dynamics .
we conclude that the effects of micromotion can be reduced by choosing ion / atom combinations with a large mass ratio and by choosing large inter - well distances .
the proposed double - well system may be realised in an experiment by combining either optical traps or magnetic microtraps for atoms with ion trapping technology .
recent experiments involving a combination of ultra - cold quantum gases and trapped ions have sparked significant interest in studying their properties in the quantum regime @xcite .
these experiments explore sympathetic cooling of ions by means of clouds of cold atoms @xcite and study cold chemistry @xcite .
trapped single ions may be used to perform in situ measurements of cold atomic gases in lattice potentials @xcite , or to study the physics of impurities in one - dimensional bose gases @xcite .
the excellent controllability of trapped ions together with the near perfect state - preparation and read - out may also allow experiments in which ions _
control _ the dynamics of ultra - cold atoms .
for instance , it has been proposed that an entangling quantum gate operation could be performed on a single trapped atom and ion by means of a controlled collision @xcite . here
, the spin dependence of the scattering cross section can be used to obtain a state - dependent collisional phase shift , leading to the desired quantum logic @xcite . in a recent paper
we have shown that a similar controllability should arise when a single trapped ion is placed in between a double - well atomic josephson junction . here
, the spin of the ion would control the tunnelling rate of atoms between the wells via the state - dependent short - range interactions of the atoms and ion @xcite .
since the ion could in this way control many - body dynamics , mesoscopic entanglement between the atomic matter wave and the spin of the ion may be created .
the interplay between the spin - dependent tunnelling and the inter - atomic interactions could also result in superpositions of quantum self - trapping @xcite and josephson tunnelling .
since trapped ions allow for superb experimental control , they may be better suited to investigate josephson physics than using single atomic impurities @xcite . including a single trapped ion expands on the rich dynamics of the josephson junction as described in numerous experimental and theoretical works @xcite .
furthermore , the system may be seen as a ` unit cell ' for a larger scale hybrid atom - ion quantum simulator @xcite .
constructing such a device by concatenating ion - controlled double - wells could be a natural way to combine quantum simulators in which atoms can tunnel between sites in an optical lattice @xcite with simulators employing the pseudo - spin and collective motional states of ion crystals @xcite .
in such a system , atomic bloch waves would interact with phononic excitations in the ion crystal , leading to solid state phenomena such as peierls instabilities @xcite and phonon - mediated interactions .
( color online ) we consider a setup in which an atom is trapped in a double - well potential with a single ion trapped in its center .
the quantised motional states of the ion and its internal spin state influence the atomic tunnelling rate between the wells .
these degrees of freedom can in turn be manipulated with laser light or radio frequency fields.,width=226 ] in reference @xcite we solved the atomic dynamics in the ion - controlled double - well system by assuming that the ion is pinned to the center of its trap . in this work
we investigate how the dynamics of the ion changes the picture obtained in ref .
@xcite . to this end
, we consider a setup in which an atom is trapped in a double - well potential with a single ion trapped in its center as shown in fig .
[ fig : junction ] .
we will numerically solve the combined atom - ion dynamics in one dimension ( 1d ) in terms of quantum defect theory ( qdt ) @xcite .
this enables us to investigate the possibility of using non - classical states of the ion motion to control the tunnelling and the effect of imperfect ground - state cooling on the dynamics .
we also address the role of micromotion - the fast oscillating motion of ions caused by a time - dependent paul trap - on the tunnelling dynamics .
this effect has been shown to significantly change the dynamics of the proposed atom - ion quantum gate @xcite , as described in ref .
@xcite , leading to slower gates or requiring additional control pulses .
micromotion has also been shown to limit attainable temperatures for ions that are sympathetically cooled by atoms @xcite . here
we show that the micromotion causes difficulties in state - preparation and complicates the atomic tunnelling for large tunnelling rates , i.e. small inter - well separations .
additionally , we find that using an atom - ion combination with a large mass ratio , with the ion being the heavier particle , allows to overcome such difficulties .
this conclusion is in line with a recent classical analysis studying attainable temperatures in atom - ion sympathetic cooling @xcite .
the paper is structured as follows : in section [ sec_static ] we will discuss the solutions to the atomic double - well in the presence of a static ion . in section [ sec_moving ] , we will calculate how the situation changes when the ion dynamics are taken into account .
we will show that the motional state of the ion is coupled to the atomic tunnelling rate such that the tunnelling can be controlled by engineering ( non - classical ) ionic states of motion .
we will also analyze the effect of imperfect ion cooling . in section [ sec_mmotion ]
we will analyze the problem in the presence of a time - dependent trapping potential for the ion and the effect of micromotion .
then , we discuss possible experimental implementations and draw conclusions in sec .
[ sec : expimp ] and sec .
[ sec : concl ] , respectively .
finally , in the appendix we shall solve the double well problem in three dimensions ( 3d ) and demonstrate that the 1d calculation results in the same physical picture . |
transition metal nitrides are of great technological and fundamental importance because of their strength and durability as well as their useful optical , electronic , and magnetic properties.@xcite by incorporating b , c , n , or o atoms into the interstitial sites of transition metals with highly valence electron density ( ved ) , the compounds of transition metal with highly hardness and strong incompressibility can be synthesized by means of high pressure experiments.@xcite for these reasons , most recently , three transition metal dinitrides , ptn@xmath0,@xcite irn@xmath0,@xcite and osn@xmath0,@xcite have been successfully synthesized under extreme conditions of pressure and temperature .
very large bulk moduli observed in experiments make them be candidates of superhard materials , which could be used in cutting tools and wear - resistant coatings .
meanwhile , many theoretical investigations have been performed to explore their structures , for that the crystal structure is an important prerequisite of understanding their physical properties .
the search for the structural forms of ptn@xmath0 , osn@xmath0 and irn@xmath0 with the lowest energy and mechanical stability usually considers hypothetic structures of orthorhombic , tetragonal , hexagonal and cubic crystal lattice .
the structures of ptn@xmath0 and osn@xmath0 are indexed as cubic ( pyrite structure , @xmath4)@xcite and orthorhombic ( marcasite structure , @xmath5).@xcite while it is very difficult to determine the structure of irn@xmath0 in experiment because of the large intensity ratio ( 1:100 ) of x - ray diffraction peaks between nitride and pure metal .
the first - principles calculations have shown that irn@xmath0 is monoclinic ( cosb@xmath0 structure , @xmath6).@xcite as a neighbor of noble transition metal os , the diboride of re has been successfully synthesized,@xcite however the dinitride of re to date has not been reported experimentally .
theoretically , zhao _ et al _ suggested that ren@xmath7 with @xmath8 structure is thermodynamically stable at ambient conditions and up to 76 gpa.@xcite here , we readdress structural and elastic properties of ren@xmath7 and find that a new phase with @xmath1 symmetry is more stable than that with @xmath8 symmetry .
all calculations are performed by the castep code@xcite using _ ab initio _
pseudopotentials based on dft with the exchange - correlation functional of ceperley and alder as parameterized by perdw and zunger ( lda - capz).@xcite
all the possible structures concerned are optimized by the bfgs algorithm ( proposed by broyden , fletcher , goldfarb , and shannon),@xcite which provides a fast way of finding the lowest energy structure and supports cell optimization in the castep code .
the interaction between the ions and the valence electrons is described by using vanderbilt s supersoft pseudopotential@xcite .
we used a plane wave cutoff of 310 ev and a brillouin zone sampling grid spacing of 2@xmath9@xmath10 0.04 @xmath11 .
pseudoatomic calculations are performed for n 2@xmath122@xmath13 and re 5@xmath125@xmath145@xmath156@xmath12 . in the geometrical optimization ,
all forces on atoms are converged to less than 0.002 ev / , all the stress components are less than 0.02 gpa , and the tolerance in self - consistent field ( scf ) calculation is 5.0@xmath1010@xmath16 ev / atom .
relaxation of the internal degrees of freedom is performed at each unit cell compression or expansion .
the standard method is used to calculate the elastic constants , in which the second derivatives of the internal energy of a crystal are determined as a function of properly chosen lattice distortions describing strains . in practice , by applying small elastic strains , the elastic constants are obtained from the change of energy or stress .
elastic moduli are given by using the elastic constants according to the voigt - reuss - hill ( vrh ) approximation.@xcite the hill average , in general , is selected as the estimation of bulk modulus and shear modulus .
to search for the most stable structure of ren@xmath7 , we consider monoclinic , orthorhombic , tetrahedral , hexagonal , and cubic structures with a re : n stoichiometry of 1:2 , including these initial structures of aute@xmath7 ( @xmath17 , no .
12 ) , cosb@xmath0 ( @xmath18 , no .
14 ) , ges@xmath7 ( @xmath19 , no .
43 ) , cacl@xmath7 ( @xmath5 , no .
58 ) , osb@xmath7 ( @xmath20 , no .
59 ) , @xmath21-reo@xmath7 ( @xmath1 , no .
60 ) , @xmath22 ( 8@xmath23 wycknoff sites , no .
61 ) , @xmath24 ( 4@xmath23 wycknoff sites , no .
62 ) , resi@xmath7 ( @xmath25 , no .
71 ) , sis@xmath7 ( @xmath26 , no .
72 ) , cu@xmath7sb ( @xmath27 , no .
129 ) , rutile ( @xmath8 , no .
136 ) , mosi@xmath7 ( @xmath28 , no .
139 ) , @xmath29 ( 1@xmath30 and 2@xmath31 wycknoff sites , no 164 ) , crsi@xmath7 ( @xmath32 , no .
180 ) , fe@xmath7p ( @xmath33 , no .
189 ) , alb@xmath7 ( @xmath34 , no .
191 ) , reb@xmath7 ( @xmath35 , no .
194 ) , pyrite ( @xmath4 , no .
205 ) , fluorite ( @xmath36 , no .
225 ) , and niti@xmath7 ( @xmath37 , no .
227 ) . combining total energy with elastic constants analysis , four competing structures ,
i.e. , @xmath1 , @xmath8 , @xmath20 , and @xmath36 , satisfy the mechanical stability critetion,@xcite in which the latter three phases have been discussed by zhao et al.@xcite the equilibrium volume _ _ v__@xmath38 and bulk modulus b@xmath38 are determined by fitting the total energy as a function of volume to the 3rd - order birch murnaghan equation of state ( eos).@xcite our calculated equilibrium parameters ( see table i ) show that the rutile ( @xmath8 ) and the _ pbcn _ phases have larger equilibrium volumes than the fluorite ( _ fm@xmath39 m _ ) phase . from the calculated relative enthalpy @xmath40h
, one can see that the _ pbcn _ phase is the most stable structure at zero pressure . in the optimized _ pbcn _ structure ,
four rhenium atoms occupy 4__c _ _ wycknoff sites ( @xmath41=0.1191 ) and eight n atoms hold 8__d _ _ sites ( @xmath42=0.2663 , @xmath41=0.3526 , @xmath43=0.0871 ) .
the nearest distance between re ( n ) and n atoms is 1.955 ( 2.644 ) .
the orthorhombic @xmath1 structure possesses a structure characterized by zigzag chains of re atoms propagating along the @xmath23 axis of the unit cell with bond length and bond angle of 2.794 and 122.4@xmath44 , respectively . in order to determine the structural phase transition between competing structures , the enthalpy @xmath45 of each phase of ren@xmath0
is computed . the enthalpy difference per formula unit ( f.u . ) as a function of pressure is plotted in fig .
1 . we do not observe structural phase transition from @xmath1 phase to others up to 100 gpa .
it is interesting to notice that @xmath1 structure has been observed in experiment being the ground state of reo@xmath7.@xcite ved is defined as the total number of valence electron divided by volume per formula unit , which is an important factor when searching the superhard materials .
the valence electron shell of re is 6s@xmath46 and 5d@xmath47 , and that of n is 2s@xmath46p@xmath48 so that the total number of valence electrons is 17 for ren@xmath7 .
ved obtained in four phases ( table i ) is slightly lower than that of re metal ( 0.4761 electrons / @xmath48 ) due to the adding of n atoms .
high ved is good to resist the fracture , which contributes to high bulk pressure of ren@xmath7 ( table i ) , predicting that ren@xmath0 could be considered as a candidate of a superhard material .
the elastic constants of @xmath1 structure obtained from our calculations ( _ _ c__@xmath49=365 gpa , _ _
c__@xmath50=553 gpa , _ _
c__@xmath51=610 gpa , _ _
c__@xmath52=138 gpa , _ _
c__@xmath53=230 gpa , _ _
c__@xmath54=83 gpa , _ _
c__@xmath55=307 gpa , _ _
c__@xmath56=232 gpa , _ _
c__@xmath57=242 gpa ) satisfy the mechanical stability requirement well , implying that the orthorhombic @xmath1 structure is mechanically stable . as is concluded above
, ren@xmath0 has a very high bulk modulus , and therefore can be regarded as a candidate of a superhard material .
however , a high bulk modulus by itself does not directly lead to high hardness .
in general , both high bulk modulus and high shear modulus are indispensable to high hardness .
therefore , the shear modulus must be taken into account when searching for new superhard materials .
the bulk modulus represents the resistance to fracture , while the shear modulus represents the resistance to plastic deformation .
moreover , two other factors are important for technological and engineering applications : young s modulus _ e _ and poisson s ratio @xmath58 .
young s modulus , defined as the ratio between stress and strain , is used to provide a measure of stiffness of the solid .
the larger the value of _ e _ , the stiffer the material .
the young s modulus and posson s ratio for an isotropic material are given by @xmath59 , @xmath60 , respectively,@xcite where b and g represent hill average bulk modulus and shear modulus .
the calculated bulk modulus _
, shear modulus _
g _ , and young s modulus
_ e _ of ren@xmath0 with @xmath1 structure are 334 gpa , 127 gpa , and 337 gpa , respectively , of which the bulk modulus of ren@xmath0 is slightly smaller than that of re metal ( 360 gpa ) .
it is worthy to note that the bulk modulus obtained from the elastic constants agrees well with the one obtained through the fit to the 3rd - order birch murnaghan eos ( b@xmath38 ) , providing a consistent estimation of the compressibility for ren@xmath0 .
ratio @xmath58 reflects the stability of a crystal against shear
. this ratio can formally take values between -1 and 0.5 , which corresponds respectively to the lower limit where the material does not change its shape , or to the upper limit when the volume remains unchanged .
poisson s ratio of @xmath1 structure is 0.3319 , which is larger than 0.3 , indicating that there exists small volume change during elastic deformation .
also possion s ratio provides more information on the characteristics of the bonding forces .
it has been proved that @xmath58=0.25 is the lower limit for central - force solids and 0.5 is the upper limit , which corresponds to infinite elastic anisotropy .
the high value @xmath58 for ren@xmath7 means that the central interatomic forces are central.@xcite in order to predict the brittle and ductile behavior of solids , pugh introduced the ratio of the bulk modulus to shear modulus of polycrystalline phases.@xcite a high ( low ) _
value is associated with ductility ( brittleness ) . the critical value which separates ductile and brittle materials is about 1.75 .
the significant difference obtained between bulk and shear modulus indicates the higher ductility of ren@xmath0 .
now , we turn to discuss the elastic anisotropy of ren@xmath0 , which is estimated by the universal anisotropy index @xmath61.@xcite the discrepancy of @xmath62 from zero determines the extent of single crystal anisotropy and accounts for both the shear and the bulk contributions . the obtained @xmath62 value is 1.252 , indicating that ren@xmath7 has a large elastic anisotropy . additionally , fractional axis compression as a function of pressure are plotted in fig .
2 . the curves presented in this figure clearly exhibit that : a bigger difference in fractional axis compression values exists for @xmath1 phase , implicating that there is a evident structure deformation with increasing pressure ; there is a transition point for compression along @xmath63- and @xmath23-axis with increasing pressure , that is , @xmath63-axis is more difficult to be compressed than @xmath23-axis at 38 gpa above .
this abnormality for compression along different axis can foretell that there is a structural phase transition at about 40 gpa .
further analysis of the change of the structural parameters ( atomic coordination , bond length , et al . ) shows that from 40 to 45 gpa fractional coordinations @xmath42 and @xmath43 in 8@xmath31 ( @xmath42 , @xmath41 , @xmath43 ) sites ( n atoms ) change much , yielding the great change of the nearest distance between n atoms ( @xmath64 ) . in detail , at 40 gpa , @xmath42 and @xmath43 are 0.2184 and 0.0752 , respectively .
however , @xmath42 and @xmath43 reduce to 0.1385 and 0.0349 at 45 gpa , respectively .
correspondingly , the @xmath64 reduces from 2.359 at 40 gpa to 1.432 at 45 gpa .
further , the debye temperature @xmath65 which correlates with many physical properties of materials , such as specific heat , elastic constants , and melting temperature is given .
the debye temperature @xmath65 can be calculated by the equation @xmath66^{1/3 } [ \frac{2}{3}(\frac{\rho}{g})^{3/2}+\frac{1}{3 } ( \frac{\rho}{b+4g/3})^{3/2}]^{-1/3}$],@xcite where @xmath67 is plank s constant , @xmath68 is boltzmann s constant ,
@xmath69 is avogadro s number , @xmath70 is density , @xmath71 is the molecular weight and @xmath72 is the number of atoms in the molecule .
the calculated @xmath65 value of 735 k in @xmath1 is comparative to the values of 866.4 @xmath73 in reb@xmath0@xcite , 850 @xmath73 in reo@xmath7@xcite and 691 @xmath73 in osn@xmath0.@xcite employing the correlation between the shear modulus and vickers hardness reported by teter @xcite for a wide variety of hard materials , the indentation hardness of ren@xmath0 can be estimated to be approximately 17.1gpa , which is comparable to that of si@xmath2n@xmath3 ( 21@xmath74 gpa).@xcite it is well known that the hardness of superhard materials should be higher than 40 gpa , indicating that ren@xmath0 is not a superhard material .
et al _ @xcite proposed that the hardness is determined by elastic constant _ _
c__@xmath52 rather than shear modulus . in our case
, we note that there is little difference between the _
_ c__@xmath52 value of elastic constant and shear modulus .
accordingly , both teter estimation and jhi s viewpoint of hardness actually achieve a consistent result for ren@xmath0 . in order to understand the nature of elastic property
, the electronic density of states is also analyzed .
the total and partial density of states ( dos ) with respect to fermi level is shown in fig .
3 . finite dos at fermi level _
n_(e@xmath75 ) of 5.035 states / ev shows that ren@xmath0 is metallic .
the electrons from re-5__d _ _ and the n-2__p _ _ states both contribute much to the dos near the fermi level .
that is , the electronic structure of ren@xmath0 is governed by strong hybridization between the re-5@xmath31 and n-2@xmath76 states , imaging that there is a covalent bond in between re and n atoms .
the existence of covalent bonding contributes to large bulk modulus , i.e. , strong incompressibility .
however , we have not found any pseudogap that observed in reb@xmath0,@xcite osb@xmath0@xcite and osn@xmath0,@xcite implying that covalent bonding in between re and n atoms is not so strong . besides , the calculated net charges on n and re atom are -0.57 e and 1.15 e , which indicates that the chemical bonding between re and n have some characteristics of ionicity .
the occurrence of ionic bonding results in the smaller shear modulus , which explains the reason of the not so large hardness in ren@xmath7 .
in conclusion , the most stable structure of ren@xmath0 is explored based on first - principles methods in the framework of density functional theory within local density approximation .
it results that orthorhombic @xmath1 structure is the most stable structure at ambient pressure .
ren@xmath0 is a metallic , super - incompressible solid and presents large elastic anisotropy .
the estimated hardness is comparative to that of si@xmath2n@xmath3 .
the structural determination and the discussion of the elastic and electronic properties may be of use in the synthesis and high pressure study of ren@xmath0 .
this work was supported by the special funds for major state basic research project of china(973 ) under grant no .
2007cb925004 , 863 project , knowledge innovation program of chinese academy of sciences , and director grants of caships .
part of the calculations were performed in center for computational science of caships and the shanghai supercomputer center . * * 1.0 cm * table i : * equilibrium lattice parameters,__v__@xmath38(@xmath48 ) , _ a _ ( ) , _
b_( ) , _ c _ ( ) , density @xmath70 ( g / cm@xmath48 ) , valence - electron density @xmath78 ( electrons / @xmath79 ) , bulk modulus _ _ b__@xmath38 ( gpa ) , and enthalpy difference per chemical formula unit ( f.u . ) @xmath80h ( ev ) .
_ _ v__@xmath38 is of per chemical f.u . | the structural , elastic , and electronic properties of ren@xmath0 are investigated by first - principles calculations with density functional theory .
the obtained orthorhombic @xmath1 structure is energetically the most stable structure at ambient pressure .
ren@xmath0 is a metallic , superincompressible solid and presents a rather elastic anisotropy .
the estimated debye temperature and hardness are 735 k and 17.1 gpa , respectively .
its estimated hardness is comparative to that of si@xmath2n@xmath3 . |
suppose that a velocity component @xmath4 is obtained at time @xmath5 along a one - dimensional cut @xmath6 of a turbulent flow .
the ensemble average @xmath7 has been subtracted , so as to have @xmath8 anywhere below .
then , the two - point velocity difference @xmath9 and average @xmath10 are defined at each scale @xmath1 as [ eq1 ] @xmath11 while @xmath9 has been studied in detail,@xcite @xmath10 has not thus far .
we are interested in @xmath10 and are to study its statistical features .
sreenivasan and dhruva@xcite studied the probability density distribution of @xmath12 in atmospheric turbulence .
the shape of the distribution was close to that of @xmath13 , regardless of the scale @xmath1 .
tatsumi and yoshimura@xcite studied the probability density distribution of @xmath12 at each scale @xmath1 in a theoretical manner , by applying a closure approximation to homogeneous isotropic turbulence .
hosokawa@xcite found an exact relation for turbulence that is homogeneous in the @xmath6 direction . since @xmath14 , where @xmath15 denotes an ensemble average , eq .
( [ eq1 ] ) yields @xmath16 through this equation , kolmogorovs@xcite four - fifth law for the longitudinal velocity at small scale @xmath1 in the inertial range @xmath17 is equivalent to @xmath18 here @xmath19 is the energy dissipation rate .
kholmyansky and tsinober@xcite confirmed eq .
( [ eq3b ] ) in atmospheric turbulence .
exact relations were also found by hill@xcite and germano.@xcite these studies tend to assume that , although @xmath20 represents motions at the scale @xmath1,@xcite @xmath12 does not depend on @xmath1 and represents energy - containing large - scale motions alone.@xcite if this were the case , @xmath21 for @xmath1 in the inertial range [ eq .
( [ eq3b ] ) ] would imply a correlation between small - scale motions in the inertial range and large - scale motions in the energy - containing range.@xcite such a correlation is of interest because still controversial is whether or not small - scale motions are statistically independent of large - scale motions.@xcite however , it is necessary to reconsider the assumption for @xmath12 .
we find that @xmath22 satisfies an exact scale - by - scale energy budget equation that involves eq .
( [ eq3b ] ) .
we also use experimental data to find that @xmath23 varies with @xmath1 in a universal manner .
thus , albeit dominated by large - scale motions , @xmath12 reflects motions at the scale @xmath1.@xcite we accordingly propose an assumption that @xmath12 represents motions over scales @xmath3 . since the correlation of @xmath24 with @xmath20 in eq .
( [ eq3b ] ) is attributable to motions at the scale @xmath1 , there is no more need to invoke a correlation between small- and large - scale motions .
we discuss this and other implications .
.[t1 ] parameters of grid turbulence ( gt ) , boundary layer ( bl ) , and jet ( j ) taken from ref .
: kinematic viscosity @xmath25 , mean rate of energy dissipation @xmath26 , kolmogorov velocity @xmath27 , rms velocity fluctuations @xmath28 and @xmath29 , skewness factor @xmath30 , flatness factors @xmath31 and @xmath32 , kolmogorov length @xmath33 , correlation lengths @xmath34 and @xmath35 , and reynolds number re@xmath36 for the taylor microscale @xmath37^{1/2}$ ] . [ cols="<,^,^,^,^",options="header " , ]
experimental data of grid turbulence , boundary layer , and jet are used here .
they were obtained with a wind tunnel in our recent work.@xcite we outline the experiments and the data .
their parameters are shown in table [ t1 ] .
for the grid turbulence , we set a grid across the wind tunnel .
the grid was made of rods with @xmath38m@xmath39 in cross section and @xmath40 m in separation . at @xmath41 m downstream of the grid ,
a hot - wire anemometer was used to measure the streamwise velocity @xmath42 and the spanwise velocity @xmath43 . here
@xmath44 is the average , whereas @xmath45 and @xmath43 are temporal fluctuations .
the data are long so that their statistical significance is high .
also at @xmath46 m and @xmath47 m , we measured @xmath42 .
this measurement was not simultaneous with the above measurement but was under the same condition .
the temporal fluctuations @xmath45 and @xmath43 measured at position @xmath48 are converted into the spatial fluctuations @xmath49 and @xmath50 at time @xmath5 , by using taylor s frozen - eddy hypothesis @xmath51 since the grid turbulence was stationary in @xmath52 and was decaying in @xmath48 , it is homogeneous in @xmath6 and is decaying in @xmath5.@xcite the data at @xmath46 m and @xmath47 m are used to estimate @xmath53 and so on at the time @xmath5 corresponding to @xmath41 m .
the data of the boundary layer and jet were obtained in the same manner as for the grid turbulence .
for the boundary layer , roughness was set over the floor of the wind tunnel .
the measurement was done in the log - law region at a streamwise position @xmath48 where the boundary layer had been well developed . for the jet
, a nozzle was set within the wind tunnel .
the measurement was done where the flow had become turbulent .
since these measurements were under stationary conditions , we use taylor s hypothesis to obtain spatial fluctuations that are homogeneous in the @xmath6 direction [ eq .
( [ eq4 ] ) ] . to explore common features
, we compare the boundary layer and jet with the grid turbulence .
the comparison is based on the correlation length @xmath34 ( table [ t1 ] ) , which represents large scales and also serves as the typical size of energy - containing eddies .
usually , statistics of @xmath20 and @xmath54 are discussed by assuming that @xmath20 and @xmath54 represent motions at the scale @xmath1.@xcite to be exact , we also have to consider motions at scales @xmath55 .
their contributions to @xmath20 and @xmath54 are nevertheless negligible as long as @xmath1 is not too large.@xcite this is because the energy distribution of a turbulent flow is a continuously and rapidly increasing function of the scale @xmath1 up to about the correlation length @xmath34 .
then , as long as @xmath1 is not too large for @xmath20 and @xmath54 to be safely assumed as representatives of motions at the scale @xmath1 , we assume @xmath12 and @xmath56 as representatives of motions over scales @xmath3 .
if this is the case , @xmath12 and @xmath56 reflect motions at the scale @xmath1 , albeit dominated by motions at the larger scales , @xmath57 .
those at scales @xmath55 also contribute to @xmath12 and @xmath56 , but they are expected to be negligible as long as they are negligible in @xmath20 and @xmath54 . our assumption is justified by using the experimental data to compare @xmath58 with @xmath59 . here ,
@xmath60 which defines the motions over scales @xmath3 because those at scales @xmath55 have been smoothed away .
the results are shown in figs .
[ f1](a ) , [ f2](a ) , and [ f3](a ) .
we observe @xmath61 for @xmath62 , where it is safe to assume @xmath12 as a representative of motions over scales @xmath3 .
this is not the case for @xmath63 , where motions at scales @xmath55 are no longer negligible in @xmath58 as well as in @xmath64 and accordingly these two are constant .
our assumption is self - consistent .
for example , in the limit @xmath65 , @xmath12 and @xmath56 reduce to @xmath13 and @xmath66 that represent motions over all the scales .
the scales of motions could be also defined with wave number @xmath67 .
we have [ eq6 ] @xmath68 e_u(k ) dk , \\ & \langle u_+^2(r ) \rangle \propto \int^{\infty}_{0 } [ 1+\cos(kr ) ] e_u(k ) dk,\end{aligned}\ ] ] where @xmath69 is the energy spectrum .
this is consistent with our assumption because @xmath70 is maximal at @xmath71 and falls to zero at @xmath72 , where @xmath73 is maximal.@xcite however , being localized in the @xmath6 space , @xmath10 or @xmath9 is not localized in the @xmath67 space . for the scales , our definition based on eq .
( [ eq5 ] ) is preferable .
phenomenologically , @xmath20 and @xmath54 represent the deformation and rotation of an eddy with size @xmath1 , respectively , if @xmath1 does not exceed the typical size of energy - containing eddies , @xmath34 .
this is just the @xmath1 range for @xmath20 and @xmath54 to represent motions at the scale @xmath1 [ figs .
[ f1](a ) , [ f2](a ) , [ f3](a ) ] .
likewise , we assume @xmath12 and @xmath56 as the advection of an eddy with size @xmath1 .
the advection is due to sweeping by nearby eddies with sizes @xmath74 , being consistent with the above assumption that motions over scales @xmath3 are represented by @xmath12 and @xmath56 . between @xmath10 and @xmath9 ,
exact statistical relations exist if the turbulence is homogeneous in the @xmath6 direction.@xcite we have already mentioned @xmath75 [ eq . ( [ eq2 ] ) ] . in addition , since @xmath76 , eq . ( [ eq1 ] )
yields [ eq7 ] @xmath77 the same relations exist between @xmath78 and @xmath79 . for each @xmath1 ,
the mean total energy @xmath80 is decomposed into the mean energies @xmath58 and @xmath64 [ eq . ( [ eq7a ] ) ] .
although @xmath12 does not correlate with @xmath20 [ eq .
( [ eq7b ] ) ] , @xmath24 could correlate with @xmath20 [ eq .
( [ eq2 ] ) ] .
this correlation is through motions at the scale @xmath1 , if we ignore motions at scales @xmath55 . the relations in eqs .
( [ eq2 ] ) and ( [ eq7 ] ) are not restricted to the homogeneous turbulence .
the same relations exist in turbulence that is isotropic throughout the scales.@xcite we nevertheless focus on the homogeneous turbulence , which is more common than the isotropic turbulence .
for example , an experiment is usually under a stationary condition .
the measured temporal fluctuations are converted by taylor s hypothesis into spatial fluctuations that are homogeneous in the @xmath6 direction [ eq .
( [ eq4 ] ) ] .
the scale - by - scale energy budget is studied in terms of @xmath22 .
we consider turbulence that is homogeneous and isotropic throughout the scales and also is decaying freely in time , for which an exact equation is known:@xcite [ eq8 ] @xmath81 \nonumber \\ & \qquad \qquad \qquad \qquad = \frac{15 \nu}{2r } \frac{\partial \langle u_-^2(r ) \rangle}{\partial r } -\frac{5 \langle u_-^3(r ) \rangle}{4r}.\end{aligned}\ ] ] by substituting eqs .
( [ eq2 ] ) and ( [ eq7a ] ) into eq .
( [ eq8a ] ) and also by rearranging the terms in the left - hand side , we obtain @xmath82 \nonumber \\ &
\qquad \qquad = -\frac{30 \nu}{r } \frac{\partial \langle u_+^2(r ) \rangle}{\partial r } + \frac{15 \langle u_+^2(r ) u_-(r ) \rangle}{r}.\end{aligned}\ ] ] thus , @xmath22 satisfies an exact scale - by - scale equation .
this would not exist if @xmath12 were a representative of large - scale motions alone .
it does represent motions at scales @xmath3 . to clarify the meanings of eqs .
( [ eq8a ] ) and ( [ eq8b ] ) , their left - hand sides are rewritten by using the relation @xmath83 which is from the asymptote @xmath84 for @xmath85 .
the results are [ eq10 ] @xmath86 = \frac{15 \nu}{2r } \frac{\partial \langle u_-^2(r ) \rangle}{\partial r } -\frac{5 \langle u_-^3(r ) \rangle}{4r } , \ ] ] @xmath87 = -\frac{30 \nu}{r } \frac{\partial \langle u_+^2(r ) \rangle}{\partial r } + \frac{15 \langle
u_+^2(r ) u_-(r ) \rangle}{r } .\ ] ] these equations offer different descriptions of the same energy budget . by using eqs .
( [ eq2 ] ) and ( [ eq7a ] ) , the left - hand side of eq .
( [ eq10b ] ) and the individual terms in its right - hand side are able to be obtained one - by - one from those of eq .
( [ eq10a ] ) .
thus , each two of these corresponding terms have exactly the same values .
since eq .
( [ eq10a ] ) describes the budget of kinetic energy per unit time and mass in terms of @xmath88 ( see below ) , it is natural to consider that eq .
( [ eq10b ] ) describes the same energy budget in terms of @xmath22 . here
@xmath88 or @xmath22 is not used as the energy itself but is used to describe its budget . although @xmath88 and @xmath22 behave differently [ figs .
[ f1](a ) , [ f2](a ) , and [ f3](a ) ] , they are statistically dependent on each other through eqs .
( [ eq2 ] ) and ( [ eq7a ] ) . for the grid turbulence , fig .
[ f1](c ) shows values of the terms of eqs .
( [ eq10a ] ) and ( [ eq10b]).@xcite their corresponding terms are actually not distinguishable at all . the left - hand side of eq .
( [ eq10 ] ) is just a weighted average of @xmath89 or @xmath90 over @xmath91 , which describes the mean rate of the total loss of the energy over scales @xmath3 .
it reduces to @xmath92 as @xmath65 and to @xmath93 as @xmath85 . the first term in the right - hand side of eq .
( [ eq10 ] ) describes the mean rate of the energy dissipation into heat over scales @xmath3 .
it reduces to @xmath94 as @xmath65 and to @xmath93 as @xmath85 .
these two are balanced with the second term , which describes the mean rate of the energy transfer across @xmath1 from the larger to the smaller scales.@xcite the transfer term for @xmath22 involves @xmath9 [ eq . ( [ eq10b ] ) ] , which is phenomenologically related to the typical time scale @xmath95 for the energy transfer across the scale @xmath1 ( sec . [ s5 ] ) . between each two of the corresponding terms of eqs .
( [ eq10a ] ) and ( [ eq10b ] ) , there is a sign difference .
this difference is associated with the fact that @xmath96 tends to increase with @xmath1 while @xmath24 tends to decrease with @xmath1 [ figs .
[ f1](a ) , [ f2](a ) , and [ f3](a ) ; see also eq .
( [ eq7a ] ) ] . when the scale @xmath1 is small enough to lie in the inertial or dissipative range , the left - hand side of eq .
( [ eq10a ] ) is able to be replaced by @xmath26 : [ eq11 ] @xmath97 this equation in turn applies to small scales of any turbulence that is isotropic over these small scales.@xcite if the turbulence is also homogeneous in the @xmath6 direction , eq .
( [ eq10b ] ) is likewise rewritten as @xmath98 for the boundary layer and jet , figs .
[ f2](c ) and [ f3](c ) show values of the terms in the right - hand sides of eqs . (
[ eq11a ] ) and ( [ eq11b]).@xcite within the inertial range , the right - hand sides of eqs .
( [ eq11a ] ) and ( [ eq11b ] ) are dominated by the second terms .
they correspond to eqs .
( [ eq3a ] ) and ( [ eq3b ] ) .
that is , while the transfer term for @xmath88 in eq .
( [ eq11a ] ) corresponds to kolmogorovs@xcite four - fifth law ( [ eq3a ] ) , the transfer term for @xmath22 in eq .
( [ eq11b ] ) corresponds to hosokawas@xcite relation ( [ eq3b ] ) .
to understand how the mean rates of energy transfer @xmath99 and @xmath100 in eqs .
( [ eq10 ] ) and ( [ eq11 ] ) are determined as observed in figs .
[ f1](c ) , [ f2](c ) , and [ f3](c ) , we study fluctuations of @xmath20 and @xmath101 .
figure [ f4](a ) shows the probability density distribution of @xmath20 in the grid turbulence . if @xmath102 , the distribution is skewed with a long tail in the negative side .
this tail explains the observed positive value of @xmath103 .
the value vanishes if @xmath63 , where the distribution of @xmath20 is symmetric .
figure [ f4](b ) shows the average @xmath104 conditioned on @xmath20 in the grid turbulence . if @xmath62 , @xmath104 is greater at @xmath105 than it is at @xmath106 .
this asymmetry overbalances the negative skewness of @xmath9 and explains the observed positive value of @xmath100 .
the value vanishes if @xmath63 , where @xmath104 is symmetric .
the same @xmath1 dependences are observed for the boundary layer in fig .
[ f5 ] with @xmath107 and for the jet in fig .
[ f6 ] with @xmath108 . besides the asymmetry , @xmath104 has features that are not common to the grid turbulence , boundary layer , and jet [ figs .
[ f4](b ) , [ f5](b ) , and [ f6](b ) ] .
the configuration for turbulence production affects large - scale motions , which in turn affect @xmath24 at any @xmath1 . however , not affected is the mean rate of energy transfer @xmath100 because @xmath24 correlates with @xmath20 only through motions at the scale @xmath1 ( sec .
[ s3 ] ) .
the local rate of energy transfer is known to fluctuate markedly.@xcite this fluctuation could be related to those of @xmath109 and @xmath110 , although they are not exactly identical to the local rate .
phenomenologically , when an eddy typical for the size @xmath1 deforms and breaks apart into the smaller daughter eddies , the deformation energy @xmath64 of the mother eddy is transferred to those of the daughter eddies . since the time scale is @xmath95 ,
this offers a conceptual explanation for the mean rate of energy transfer @xmath103 .
the mother eddy also has energy @xmath58 for the advection , which is also transferred to those of the daughter eddies . with the time scale @xmath95 , this explains the mean rate in terms of @xmath100 .
these two explanations are equivalent if we consider eddies with all the sizes , since the advection @xmath10 of an eddy reflects the deformations @xmath9 of the surrounding larger eddies . in fact
, we have @xmath111 .
the @xmath1 dependences of @xmath112 and @xmath113 are distinct from those of @xmath114 and @xmath115 .
for example , while @xmath64 varies by several orders , @xmath58 varies only from @xmath116 to @xmath80 [ figs .
[ f1](a ) , [ f2](a ) , and [ f3](a)].@xcite this range is too narrow to exhibit a power - law scaling @xmath117 such as those known for @xmath114.@xcite the @xmath1 dependences of @xmath112 and @xmath113 are also subject to large - scale motions that depend on the configuration for turbulence production . for example , although exactly isotropic turbulence exhibits @xmath118 at all @xmath1 , our grid turbulence exhibits some variation of @xmath119 . there still exist features that do not depend on the flow configuration , as demonstrated below by using long experimental data .
figure [ f1](b ) shows the flatness factors in the grid turbulence , e.g. , @xmath120 . with a decrease in @xmath1 ,
those of @xmath20 and @xmath54 increase as a result of the small - scale intermittency.@xcite those of @xmath12 and @xmath56 are constant within @xmath121% .
the reason is the predominance of large - scale motions .
in fact , @xmath122 these asymptotic values are the same if @xmath123 , which is close to the case of the grid turbulence ( table [ t1 ] ) .
however , in more detail , the flatness factors of @xmath12 and @xmath56 vary with the scale @xmath1 , exhibiting maxima at @xmath124 and minima at @xmath125 . albeit of small amplitude
, these variations are still above the statistical uncertainties that are seen as rapid fluctuations at @xmath126 .
the same variation exists for the flatness factor of @xmath56 in the boundary layer and jet [ figs .
[ f2](b ) and [ f3](b)].@xcite for the observed variations of the flatness factors , we attempt an explanation , which is not complete but would serve as a guide for future studies .
when @xmath1 is not too large , @xmath12 represents the advection of an eddy with size @xmath1 , which is due to sweeping by nearby eddies .
if each of them has the same sweeping direction , the corresponding @xmath10 value tends to lie at the tail of the @xmath10 distribution .
since eddies with sizes @xmath127 do not contribute to the sweeping , an increase in @xmath1 shortens the above tail .
the flatness factor becomes small . with a further increase in @xmath1 , eddies with sizes @xmath74
become sparse .
the @xmath10 distribution becomes enhanced at @xmath128 , i.e. , cases of no eddies .
the flatness factor becomes large .
finally , when @xmath1 exceeds the typical size @xmath34 of energy - containing eddies , their random relative motions become dominant .
the flatness factor tends toward the gaussian value of 3 [ eq .
( [ eq12 ] ) ] .
the velocity averages @xmath12 and @xmath56 represent motions over scales @xmath3 , as long as @xmath1 is not too large for motions at scales @xmath55 to be negligible , i.e. , for the velocity differences @xmath20 and @xmath54 to be safely assumed as representatives of motions at the scale @xmath1 .
phenomenologically , while @xmath20 and @xmath54 represent the deformation and rotation of an eddy with size @xmath1 , @xmath12 and @xmath56 represent the advection of such an eddy . the motions over scales @xmath3 are defined by @xmath129 , i.e. , velocity smoothed over the scale @xmath1 [ eq . ( [ eq5 ] ) ] .
when @xmath1 is not too large , we observe @xmath61 [ figs .
[ f1](a ) , [ f2](a ) , and [ f3](a ) ] .
we have found an exact scale - by - scale energy budget equation for @xmath22 [ eq .
( [ eq10b ] ) or ( [ eq11b ] ) ] , which involves the relation found by hosokawa@xcite [ eq . ( [ eq3b ] ) ] as the mean rate of energy transfer across the scale @xmath1 .
this equation describes the same energy budget as the equation for @xmath88 [ eq .
( [ eq10a ] ) or ( [ eq11a ] ) ] . between them ,
the corresponding terms have exactly the same values .
we have also found that the flatness factors of @xmath12 and @xmath56 vary with @xmath1 in a manner that is independent of the configuration for turbulence production [ figs .
[ f1](b ) , [ f2](b ) , and [ f3](b ) ] .
thus , albeit subject to large - scale motions that depend on the flow configuration , @xmath10 and @xmath78 have universal features .
they are localized around @xmath132 , where @xmath133 is some unit scale . while @xmath134 represents motions over scales @xmath135 , @xmath136 represents motions at the scale @xmath137 .
if @xmath138 is not too large , the transform @xmath139 is analogous to the velocity difference @xmath9.@xcite the transform @xmath140 is analogous to the velocity average @xmath10 .
there exist various wavelets based on various definitions of the scales . those for haar s wavelet@xcite are the same as defined for the present study [ eq .
( [ eq5 ] ) ] .
tatsumi and yoshimura@xcite proposed a closure approximation that @xmath10 is statistically independent of @xmath9 at each scale , which is interpreted as an approximation of statistical independence between the advection @xmath12 and deformation @xmath20 of eddies with each size @xmath1 .
this is exact in many cases [ eq . ( [ eq7b ] ) ] .
an exception is the correlation of @xmath22 with @xmath9 studied here [ eq . ( [ eq2])].@xcite there is a controversy as to whether or not small - scale motions in the dissipative and inertial ranges are statistically independent of large - scale motions in the energy - containing range.@xcite to address this issue , @xmath141 at small @xmath1 is often compared with @xmath142 or @xmath143 , by assuming that they represent the large - scale motions .
being actually a representative of motions over scales @xmath3 , @xmath142 could correlate with @xmath141 through motions at the scale @xmath1 [ eq . ( [ eq2 ] ) ] .
the situation is the same for @xmath143 , which corresponds to @xmath142 for @xmath65 and hence could correlate with @xmath141 .
such correlations are spurious as those between small- and large - scale motions .
in particular , the large - scale motions are not responsible to @xmath21 for @xmath1 in the inertial range [ eq .
( [ eq3b ] ) ] . to conclude this paper
, we underline that further studies of @xmath10 and @xmath78 are promising.@xcite for the above and other issues , @xmath10 and @xmath78 could reveal new features and offer new insights .
this is because the description of turbulence by @xmath10 and @xmath78 differs from the existing description by @xmath9 and @xmath79 , and also because @xmath10 and @xmath78 satisfy exact equations like eq .
( [ eq10b ] ) that serve as the firm theoretical basis .
we are grateful to k. hashimoto and y. kawashima for continued collaborations as well as to i. hosokawa , t. gotoh , m. takaoka , t. tatsumi , s. toh , and a. tsinober for useful comments .
this research was supported in part by kakenhi ( c ) 22540402 .
a. a. praskovsky , e. b. gledzer , m. y. karyakin , and y. zhou , `` the sweeping decorrelation hypothesis and energy - inertial scale interaction in high reynolds number flows , '' j. fluid mech .
* 248 , * 493 ( 1993 ) .
turbulence in this hypothetical space @xmath6 is still realistic .
local values of @xmath49 and @xmath50 represent local motions of the actual turbulence .
they are continuously connected up to largest scales ( ref . ) .
both of @xmath144 extend to high @xmath67 without any decay , corresponding to the contributions of motions at scales @xmath145 .
their significance depends on the @xmath1 value with respect to the shape of @xmath69 .
there could exist turbulence that is not homogeneous but is still isotropic around some specific position @xmath146 . by setting @xmath146 to be the midpoint of @xmath6 and @xmath147 and by taking ensemble averages
, we obtain the same relations as in eqs .
( [ eq2 ] ) and ( [ eq7 ] ) , although they do not necessarily hold around the other positions .
normalized by different estimates of @xmath26 to compensate for unavoidable artifacts in our experiments .
the measurement of the @xmath13 component was not so reliable at smallest scales ( ref . ) .
the grid turbulence was not so isotropic at largest scales .
then , the most reliable estimate of @xmath26 is @xmath151 in table [ t1 ] .
the right - hand sides of eqs .
( [ eq10a ] ) and ( [ eq10b ] ) are determined at each scale @xmath1 .
although @xmath24 represents motions over scales @xmath3 , any information about scales @xmath57 has been lost from @xmath152 , i.e. , differential with respect to @xmath1 .
such information has been also lost from @xmath153 because @xmath24 correlates with @xmath20 only through motions at the scale @xmath1 ( sec .
[ s3 ] ) . for these boundary layer and jet , @xmath120 is not shown .
its values for @xmath65 and @xmath85 are too different because @xmath154 is too different from @xmath155 [ table [ t1 ] , see also eq .
( [ eq12 ] ) ] , although there appears to exist the same variation as observed for the grid turbulence . | for turbulence , although the two - point velocity difference @xmath0 at each scale @xmath1 has been studied in detail , the velocity average @xmath2/2 $ ] has not thus far .
theoretically or experimentally , we find interesting features of the velocity average .
it satisfies an exact scale - by - scale energy budget equation .
the flatness factor varies with the scale @xmath1 in a universal manner .
these features are not consistent with the existing assumption that the velocity average is independent of @xmath1 and represents energy - containing large - scale motions alone .
we accordingly propose that it represents motions over scales @xmath3 as long as the velocity difference represents motions at the scale @xmath1 . |
an activity of many astrophysical sources ( active galactic nuclei , young stellar objects , galactic x - ray sources , microquasars ) is associated with the accretion . for this reason ,
the accretion onto compact objects ( neutron stars or black holes ) is the classical problem of modern astrophysics ( see , e.g. , @xcite and references therein ) . at present the analytical approach
, whose foundation was laid back in the mid - twentieth century @xcite , began to be supplanted by numerical simulations @xcite .
analytical solutions were found only in exceptional cases @xcite . it should be emphasized that last time the focus of the research has been shifted to numerical magnetohydrodynamic simulations , within which framework it has become possible to take into account the turbulent processes associated with magnetic reconnection , magnetorotational instability , etc .
however , in our opinion , some important features of the turbulent accretion can still be understood on the ground of simple analytical model . the problem of turbulent accretion and stellar wind has been discussed in various papers .
@xcite showed that the inclusion of viscosity and heat conduction allows to remove singularity at the sonic surface and leads to the appearance of weak non - rankine - hugoniot shock waves in the solutions .
@xcite examined the spatial stability of spherical adiabatic bondi accretion on to a point gravitating mass against external vortex perturbations .
@xcite also discussed the role of turbulence in a spherically symmetric accreting system . in their paper
it was shown that the sonic horizon of the transonic inflow solution is shifted inwards , in comparison with inviscid flow , as a consequence of dynamical scaling for sound propogation in accretion process .
@xcite formulated a model of spherically symmetric accretion flow in the presence of magnetohydrodynamic turbulence .
in this paper we show on the ground of the simple model how the turbulence affects the structure of the spherically symmetric accretion . in the first part
, we formulate the basic equations of ideal steady - state axisymmetric hydrodynamics , which are known to be reduced to one second - order equation for the stream function .
then , in the second part , the structure of the solitary curl is discussed in detail .
finally , in the third part we consider two toy models describing axisymmetric turbulence .
it is shown that the turbulence changes mainly the effective gravity potential but not the effective pressure .
first of all , let us formulate basic hydrodynamical equations describing axisymmetric stationary flows . then , as is well - known , it is convenient to introduce the potential @xmath0 connected with the poloidal velocity @xmath1 and the number density @xmath2 as @xcite @xmath3 this definition results in the following properties * the continuity equation @xmath4 is satisfied automatically .
* it is easy to verify that @xmath5 , where @xmath6 is an area element . as seen , the potential @xmath0 is a particle flux through the circle @xmath7 . in particular ,
the total flux through the surface of the sphere of radius @xmath8 is @xmath9 . * as @xmath10 , the velocity vectors @xmath11 are located on the surfaces in this case , three conservation laws for energy @xmath12 , angular momentum @xmath13 , and the entropy @xmath14 can be formulated as @xmath15 here @xmath16 is the specific enthalpy , and @xmath17 is the gravitational potential . in what follows
we for simplicity consider the entropy @xmath18 to be constant .
then the equation for the stream function @xmath0 ( which is no more than the projection of the euler equation onto the axis perpendicular to the velocity vector @xmath11 ) looks like ( cf .
@xcite ) @xmath19 where @xmath20 .
this equation represents the balance of forces in a normal direction to flow lines . in partucular , for spherically symmetric flow , i.e. , for @xmath21 const , it has the solution @xmath22 in the following , we deal with the linear angular operator @xmath23 originated from eqn .
( [ gscomp ] ) .
it has eigenfunctions @xmath24 and the eigenvalues @xmath25 here @xmath26 are the legendre polynomials and the dash indicates their derivatives .
let us consider now the small disturbance of the spherically symmetric flow , so that one can write down the flux function as @xmath27 \label{gssol1new}\ ] ] with the small parameter @xmath28 . then eqn . ( [ gscomp ] )
can be linearised , while the equation for the perturbation function @xmath29 is written as @xcite : @xmath30 here @xmath31 , and @xmath32 .
this equation allows us to seek the solution in the form @xmath33 introducing now dimensionless variables @xmath34 where the @xmath35-values correspond to the sonic surface ( which can be taken from the zero approximation ) , we can write the ordinary differential equations describing the radial functions @xmath36 : @xmath37 here @xmath38 , @xmath39 , and the expansion coefficients @xmath40 , @xmath41 and @xmath42 depend on the disturbances as : @xmath43 @xmath44 @xmath45 finally , the functions @xmath46 and @xmath47 correspond to the spherically symmetric flow . for the polytropic equation of state @xmath48 we use here
they are connected by the relation @xmath49 . as to the dimensionless number density @xmath47 , it can be found from ordinary differential equation @xmath50 with the boundary conditions @xmath51 @xmath52 in region @xmath53 , @xmath54 from 1.1 to 5/3 ] as for the boundary conditions for the system of eqns .
( [ radialeq ] ) , they are taken analogous to the case of bondi - hoyle accretion @xcite : @xmath55 , \label{boundg}\ ] ] @xmath56 where @xmath57 stands for the expansion in terms of the legendre polynomials , which can be found from energy and angular momentum integral perturbations .
now let us consider in detail the internal structure of the quasi - spherical accretion with a small axisymetric perturbation localised near the axis @xmath58 .
in other words , we suppose that the angular size of the vortex is small enough ( @xmath59 ) .
the main goal of this numerical calculation is to find the disturbance function @xmath60 which gives us the possibility to determine velocity components @xmath61 and @xmath62 , i.e. , the main characteristics of the perturbed flow . to model the internal structure of the vortex , we determine @xmath63-dependent angular velocity @xmath64 in the form @xmath65 . \label{momentum integral}\ ] ] here @xmath66 gives the amplitude of considered curl and the coefficient @xmath67 ( which are free parameters of our model ) is an inversed curl width .
certainly , we assume that the perturbation is small in comparison with the main contribution of the radial accretion
. then the flow structure can be described by the system ( [ radialeq ] ) , ( [ densityeq])([boundderivg ] ) formulated in previous section . as to the expansion coefficients @xmath42 , @xmath41 and @xmath40 , they are to be determined from eqns .
( [ denergy])([dmomentum ] ) and ( [ boundg ] ) on the outer boundary of a flow @xmath68 . for our choice ( [ momentum integral ] )
the disturbances have the form @xmath69 r^2 \sin^2\theta}}{2 } + { o(\varepsilon^3 ) } , \label{energyperturb}\ ] ] @xmath70 r^2 \sin^2\theta}. \label{momentumperturb}\ ] ] as one can easily check , it gives @xmath71
. expansions ( [ denergy ] ) , ( [ squaremomentum ] ) , and ( [ dmomentum ] ) in terms of @xmath72 contain some numerical difficulties because the set of this functions is not an orthogonal one , and , even though it converges , in our case of very small curl width we can neglect just summands with numbers larger than 50 . even in some trivial cases like @xmath73 these polynomials call a number of numerical obstacles ( e.g. , bad - conditioned matrix of linear equation for coefficients @xmath42 , @xmath41 and @xmath40 , etc ) . in order to expand functions of integrals , we used the auxiliary set of chebyshev polynomials , which is orthogonal and posesses a feature of generally faster convergence . using these polynomials
, we could find all expansions with the accuracy no worse than @xmath74 . as was shown in sect .
2 , the normalized density function @xmath47 can be derived from the equation ( [ densityeq ] ) and boundary conditions ( [ densityeqbound ] ) and ( [ densityeqboundaries ] ) .
the results of numerical calculations for different polytropic indeces @xmath54 is shown on figure 1 .
in particiular , as one can see , the density is nearly constant in subsonic regime ( @xmath75 ) .
an example of the numerical calculation of perturbation function @xmath29 is shown on figure 2 .
we should stress that @xmath29 turns actually zero outside the small region near the axial curl .
this statement allows us to assume as a zero approximation that the turbulent accretion regime containing a number of curls can be considered as a set of noninteracting ones .
moving towards the accreting star , we can register that perturbation has a maximum on a radius @xmath76 , then diminishes on the distance about @xmath77 and finally rapidly rises in the vicinity of the sonic surface @xmath78 . apart from numerical solution
, we can also find the analytical asymptotic solutions in the supersonic region @xmath79 where eqn .
( [ radialeq ] ) can be rewritten as @xmath80 here we take into account that @xmath81 .
getting rid of all parameters from the right part of this equation , one can introduce a new function @xmath82 .
( [ asympteq1 ] ) can be rewritten as @xmath83 neglecting now all terms which are proportional to @xmath84 , where @xmath85 , we obtain that this equation has an universal solution independent of the boundary conditions on the outer boundary @xmath68 @xmath86 remember that the same asymptotic behavior was obtained by @xcite for homogeneously rotating flow . in region @xmath53 ,
@xmath87 , @xmath88 as was already stressed , numerical results allow us to determine @xmath89 ratio around the curl .
it is easy to show that @xmath90 where @xmath91}. \label{1}\ ] ] in our calculation we put @xmath92 , so that the function ( [ 1 ] ) is limited in area near the curl ( @xmath93 ) .
taking now into account that @xmath94 is a small parameter of our expansion and @xmath95 , one can show that for reasonable parameter @xmath94 the ratio @xmath96 in the area of vortex has an order of @xmath97 ( see figure 3 ) .
thus , we could claim that @xmath98 , and then one can neglect the all terms in navier - stokes equations that consist @xmath61 . ratio in region @xmath99 . ] on the other hand , it is necessary to stress that , according to ( [ asymptotic ] ) , we can not use our solution close enough to the origin ( @xmath100 ) . indeed , analysing the field line equation @xmath101 , we obtain that the asymptotic solution survives until @xmath102 , where @xmath103 is an initial angular size of a curl and @xmath104 is a broadening parameter of the stream line .
deriving @xmath63-component of the velocity from ( [ b1 ] ) and taking @xmath105 of zero order , we get @xmath106 assuming now that @xmath107 , one can expand equation ( [ crit ] ) in @xmath63 and neglecting all summands of @xmath108 with @xmath109 , we find @xmath110 this equation can be simply integrated , and we obtain @xmath111 hence , under the radius @xmath112 we can not use the solution ( [ asymptotic ] ) as the disturbance becomes larger than unity .
in order to keep the solution up to star surface we should demand @xmath113 where @xmath114 corresponds to the star radius .
it gives us the general condition of appliability of the approach described above . unless we can not use the method of linear expansion of grad - shafranov equation , and the turbulent flow
is to be described in another way which lies outside the consideration of current paper .
let us try now , as have been done in some papers ( see , e.g. , @xcite , @xcite ) , to describe the vortex disturbance through some effective pressure , i.e. , to calculate a small correction for the pressure function @xmath115 caused by the presence of turbulent curl .
starting with @xmath63-component of euler equation @xmath116 and neglecting all the terms containing @xmath61 , we obtain @xmath117 expanding this equation near the axis and assuming @xmath118 where @xmath119 const , we obtain @xmath120 it gives for the additional pressure @xmath121 @xmath122 where @xmath123 is an approximate size of a curl .
thus , one can conclude that the pressure @xmath121 has explicit dependence on @xmath8 , which can not be taken into consideration using any equation of state .
hence , we should include the turbulence into the consideration in another way than introducing effective turbulent pressure as often suggested by @xcite .
let us suppose now that the turbulence in the accreting matter can be described by the large number of axisymmetric vortexes with different parameters @xmath124 and @xmath125 filling all the accreting volume . within this approach one can construct two toy analytical models demonstrating how the turbulence can affect the structure of the spherically symmetric accretion .
the first model in which we neglect viscosity corresponds to classical ideal spherically symmetric bondi accretion @xcite . in this case
one can consider the following system of equations : @xmath126 following @xcite we consider polytropic equation of state @xmath127 resulting in for polytropic index @xmath128 @xmath129 here again @xmath2 ( @xmath130 ) is the number density , @xmath131 ( in @xmath132 ) is the mass of particles ( @xmath133 is the mass density ) , @xmath14 is the entropy per one particle ( dimensionless ) , is the specific enthalpy , @xmath134 ( in @xmath135 ) is the temperature in energy units , and , finally , @xmath136 ( @xmath137 ) is the sound velocity . as was demonstrated above , for a weak enough turbulence level ( [ critansw ] ) for any inividual curl one can neglect @xmath63-component of the velocity perturbation in comparison with the toroidal one @xmath62 up to the central body @xmath138 .
thus , in zero approximation one can put @xmath139 , i.e. , @xmath140 const .
this implies that , according to the angular momentum concervation law @xmath141 const , we can write down @xmath142 here @xmath64 is a smooth function of @xmath63 that can be approximately described as : @xmath143 in order to find the characteristic values of the accretion flow we have to use energy and momentum intergals conserving on stream lines .
taking into account an assertion @xmath144 , we can neglect @xmath63-component of velocity which gives @xmath145 averaging now these integrals in @xmath63 and introducing a new value @xmath146 we obtain for the averaged energy integral @xmath147 as we see , two last terms can be considered as effective gravitational potential .
futher calculations are quite similar to the classical bondi problem for the spherical flow . in other words , using another integrals of motion , i.e. , the total particle flux @xmath148 and the entropy @xmath14 , one can rewrite the energy integral ( [ e_av ] ) as @xmath149 it gives the following expression for the logarithmic @xmath8-derivative of the number density @xmath150 as for bondi accretion , this derivative has a singularity on the sonic surface @xmath151 .
this implies that for smooth transition through the sonic surface @xmath152 , the additional condition is to be satisfied : @xmath153 solving now ( [ bondieq ] ) in terms of @xmath154 in this approximation , we find @xmath155 where @xmath156 is evaluated from @xmath157 accordingly , we obtain for @xmath158 and @xmath159 ratios : @xmath160 where @xmath161 and @xmath162 correspond to the classical bondi accretion .
finally , using the definition ( [ avemomentum ] ) for @xmath163 , we can rewrite our criteria of the appliability ( [ critansw ] ) as @xmath164 . as @xmath165
, it can be finally rewritten as @xmath166 to sum up , one can conclude that the nonzero angular momentum effectively decreases the gravitational force .
in other words , the presence of the angular momentum do not allow matter to fall down as easy as in its absence .
roughly speaking , we substitute our gravitating center with one that posesses less mass .
so , in the case of bondi accretion with a small angular momentum perturbation we should modify the relations for sonic surface radius and velocity they decrease and rise respectively .
it is important to note that we can consider turbulent accretion regime as one with a modified gravity potential . for @xmath167 it this subsection
we consider stationary axisymmetric quasi - spherical flow of viscous fluid .
using the condition @xmath144 , and neglecting all the terms containing @xmath61 , we obtain for @xmath168-component of euler equation @xcite @xmath169 for viscous flow it is convenient to determine the toroidal component of the velocity @xmath170 as @xmath171 where we will use the following form for the angular velocity @xmath172 : @xmath173 .
\label{om}\ ] ] here @xmath174 is an amplitude , and @xmath175 is a square of effective angular width of an individual curl . substituting now @xmath170 into eqn .
( [ ll ] ) , we obtain @xmath176 , \label{54}\ ] ] where @xmath177 is the accretion rate remaining constant in stationary flow , and @xmath178 is a dynamic viscous coefficient which can be considered as a constant as well @xcite .
using now eqn .
( [ 54 ] ) , one can easily show that the total angular momentum of an individual vortex conserves ( @xmath179 ) .
indeed , internal friction connecting with viscosity can not change the total angular momentum of the accreting matter .
for this reason , together with ( [ accretionrate ] ) , the angular momentum @xmath180 can be rewritten as a full @xmath63-derivative .
this implies that the r.h.s .
( [ 56 ] ) intergated over @xmath63 becomes zero .
further , to determine the radial dependence of the curl amplitude @xmath181 and the squared width @xmath182 , we substitute the angular velocity @xmath183 ( [ om ] ) into ( [ ll ] ) and expand it in terms of @xmath63 near the axis , neglecting all the terms with the power more than 3 . as a result ,
we obtain two equations for @xmath181 and @xmath182 @xmath184 @xmath185 which have simple solutions @xmath186^{-2}. \end{aligned}\ ] ] introduction of a small vortex tubulence can be again treated by modifying a gravitational potential as @xmath187^{-3}.\ ] ] thus , viscosity results in increasing of the vortex width ( @xmath188 for @xmath189 corresponding to accretion ) and diminishing of the angular rotation . on the other hand ,
the role of viscosity will be small if @xmath190 for @xmath191 we return to the previous result @xmath192 const , @xmath193 .
introducing now reynolds number as @xmath194 where @xmath195 and @xmath196 are characteristic values of a flow and using expression ( [ accretionrate ] ) , we can rewrite ( [ remm ] ) as @xmath197 this implies that the role of viscosity is small for turbulent flow .
certainly , the analysis presented above allows us to take into consideration only isolated set of curls . in reality ,
dense celluar turbulent structure posesses a number of collective effects @xcite , that is to be described in another way .
the easiest method to proceed with the minimal number of additional assumptions is to choose another angular velocity profile .
as the total angular momentum of the accreting matter is suppose to be zero , we will use the following expression for the angular velocity @xmath198 here the parameter @xmath199 is to be chosen from the condition of the zero total angular momentum @xmath200 which is equivalent to @xmath201 one of its realisations can be seen on figure 4 where the dashed line shows zero angular velocity level .
configuration like this one represents the unit of celluar turbulent structure , fulfilling main requirements of its nature . to simplify our calculations , we hold @xmath199 and @xmath202 on constant values in order to get simple equation on @xmath203 . again , we expand equation ( [ 54 ] ) in terms of @xmath63 to the first order . as a result
, we obtain for @xmath204 : @xmath205 .
\label{eff3}\ ] ] in this case , the effective gravitational potential can not be derived for arbitary parameters without special functions .
it can be written as @xmath206,\ ] ] where @xmath207 choosing , for instanse , @xmath208 and @xmath209 , it gives @xmath210 .
the expression under the exponential function in ( [ eff3 ] ) is always lower than zero , so the criterion of the importance of viscosity effects can be formulated as @xmath211 which is more convinient to discuss in terms of reynolds number @xmath212 thus , for narrow vortex ( i.e. , for @xmath213 ) the viscosity effects must be taken into account .
we have shown that in the case of the vortex turbulence represented by a solitary axisymmetric curl , one can not simulate effects of the turbulence by introduction of the effective pressure which is common in recent papers ( see , e.g. , @xcite , @xcite ) . on the other hand , it is possible to take into account the vortex structure of the accreting flow by introducing an effective gravitational potential .
further , we described two analytical toy models that show how the turbulence affects the structure of the spherically symmetric flow .
in particular , it was shown that the sonic surface moves inwards because of effective diminishing of gravitational force . finally , a criterion to analyze the importance of viscosity effects in the adiabatic flow filled either by isolated or dense set of curls was formulated . anderson m. , 1989 , mnras , 239 , 19 axford w. i. & newman r. c. , 1967 , astrophys .
j. , 147 , 230 balbus s. a. & hawley j. f. , 1991 , astrophys .
j. , 376 , 214 beskin v.s .
, mhd flows in compact astrophysical objects , 2010 .
springer , berlin beskin v. s. & malyshkin l. m. , 1996 , astron .
lett . , 22 , 475 beskin v. s. & pidoprygora yu . n. , 1995 , jetp 80 , 575 beskin v. s. & pidoprygora yu .
n. , astron .
rep . , 1998 , 42 , 71 bhattacherjee j. k. & ray a. k. , 2005 , astrophys .
j. , 627 , 368 bisnovatyi - kogan g.s .
, kazhdan ya.m .
, klypin a. a. et al .
, 1979 , sov .
23 , 201 bondi h. , 1952 , mnras , 112 , 195 bondi h. & hoyle f. , 1944 , mnras , 104 , 272 brandenburg a. & sokoloff d. d. , 2002 , geophys .
fluid dyn . , 96 , 319 heyvaerts j. , 1996 , in chiuderi c. , einaudi g. eds , plasma astrophysics .
springer , berlin , p. 31 hunt r. , , 1979 , mnras , 198 , 83 kovalenko i. g. & eremin m. a. , 1998 , mnras , 298 , 861 krolik j. & hawley j. f. , 1991 , astrophys .
j. , 573 , 754 landau l. d. & lifshits e. m. fluid mechanics , 1987 .
pergamon press , oxford lipunov v. m. , astrophysics of neutron stars , 1992 , springer , heidelberg pariev v. i. , 1996 , mnras , 283 , 1264 petrich l. i. , shapiro s. , & teukolsky s. , 1988 , phys .
lett , 60 , 1781 petrich l.i .
, shapiro s. , stark r. f. , & teukolsky s. , 1989 , astrophys .
j. , 336 , 313 prokhorov m. e. , popov s. b. , khoperskov a.v .
2002 , astron .
astrophys . , 381 , 1000 ruffert m. & arnett d. , 1994 , astron .
, 346 , 861 shakura n. i. , postnov k. a. , kochetkova a. yu .
& yalmasdotter l. , 2013 , phys .
, 183 , 337 shakura n. i. , postnov k. a. , kochetkova a. yu . & yalmasdotter l. , 2012 , mnras , 420 , 216 shapiro s. & teukolsky s. , black holes , white dwarfs , and neutron stars : the physics of compact objects , 1983 , wiley , new york shcherbakov r. v. , 2008 , astrophys . j. s. , 177 , 493s toropin yu .
m. , toropina o. d. , saveliev v. v. , et al .
, 1999 , astorphys .
j. , 593 , 472 toropina o. d. , romanova m. m. & lovelace r. v. d. , 2012 , mnras , 420 , 810 | we analize the quasi - spherical accretion in the presence of axisymmetric vortex turbulence .
it is shown that in this case the turbulence changes mainly the effective gravity potential but not the effective pressure .
[ firstpage ] |
the fact that the galaxies appear to be distributed along filaments which are interconnected to form a web like structure which is often reffered as the ` cosmic web ' is one of the most striking visual feature in all the present and past redshift surveys ( e.g. , cfa , @xcite ; lcrs , @xcite ; 2dfgrs , @xcite and sdss , @xcite ) .
the analysis of filamentary patterns in the galaxy distribution has a long history dating back to a few papers in the late - seventies and mid - eighties by @xcite , @xcite , @xcite , @xcite and @xcite .
filaments are the most striking visible patterns seen in the galaxy distribution ( e.g. @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) .
a review on a number of physically motivated and statistical methods to define filaments is provided in @xcite . the percolation analysis ( eg .
@xcite , @xcite ) , the genus statistics ( eg .
@xcite ) , the minimal spanning tree ( e.g. @xcite ) , the voronoi tessellation ( @xcite , @xcite ) , the minkowski functionals ( eg . @xcite , @xcite ) and the ` shapefinders ' @xcite are some of the useful statistical tools introduced to quantify the topology and geometry of the galaxy distribution . @xcite propose to apply a marked point process to automatically delineate filaments in the galaxy distribution .
@xcite studied the intercluster filamentary network in high resolution n - body simulations of structure formation in a @xmath0cdm universe . @xcite use multiscale morphology filter technique to identify wall - like and filament - like structures in cosmological n - body simulations .
@xcite propose a skeleton formalism to quantify the filamentary structure in three dimensional density fields .
@xcite propose to apply an object point process to objectively identify filaments in galaxy redshift surveys .
@xcite propose the local dimension to locally quantify the shape of large scale structures in the neighbourhood of different galaxies in the cosmic web .
the sdss @xcite is currently the largest galaxy redshift survey . in an earlier work @xcite ( hereafter paper i ) we have analysed the filamentarity in the equatorial strips of this survey .
these strips are nearly two dimensional and we have projected the data onto a plane and analysed the resulting 2-d galaxy distribution .
we find evidence for connectivity and filamentarity in excess of that of a random point distribution , indicating the existence of an interconnected network of filaments .
we find that filaments are statistically significant upto length scales @xmath1 and not beyond @xcite .
all the structures spanning length - scales larger than this length scale are the result of chance alignments .
this is consistent with an earlier analysis by @xcite where they show that in las campanas redshift survey ( lcrs ) the largest length - scale at which filaments are statistically significant is between @xmath2 to @xmath3mpc .
further we show that the average filamentarity of the galaxy distribution depends on various physical properties of galaxies such as luminosity , colour , morphology and star formation rate @xcite .
it would be interesting to know if the statistically significant length scale of filaments also depends on different galaxy properties . in the present work we study if the statistically significant length scale of filaments depends on luminosity , colour and morphology of galaxies .
further it is also possible to measure the statistically significant length scale of filaments in mock galaxy distribution extracted from n - body simulation .
the model is currently believed to be the minimal model which is consistent with most cosmological data ( @xcite ; @xcite ; @xcite ; @xcite ; @xcite ; @xcite ) .
it would be interesting to measure the statistically significant length scale of filaments in n - body simulations of @xmath0cdm model and compare it with measured values in the sdss @xcite .
the n - body simulations primarily predict the clustering of the dark matter .
this should be contrasted with the fact surveys reveal only the bright side of the matter distribution .
galaxy formation is a complicated process and the exact relation between the distribution of the galaxies and the dark matter is not well understood .
it is now generally accepted that the galaxies are a biased tracer of the dark matter distribution ( e.g. , @xcite ; @xcite ) and on large scales one expects the fluctuations in the galaxy and the dark matter distribution to be linearly related through the linear bias parameter @xmath4 .
mock galaxy distributions with different bias can be simulated following this assumption .
the average filamentarity of the simulated galaxy distribution is found to be very sensitive to the bias parameter @xcite .
it would be interesting to know how the statistically significant length scale of filaments depends on the bias parameter . in the present work we test if the statistically significant length scale of filaments depends on bias .
earlier we find that the average filamentarity is sensitive to the area and galaxy number density of the samples @xcite(hereafter paper ii ) and this statistics can be used for a meaningful comparison between two different galaxy samples only when they have the same volume ( identical shape and size ) and galaxy number density . in the present work we would also like to test if the statistically significant length scale of filaments depends on area and galaxy number density of the samples .
finally we simulate segment cox process @xcite with different segment lengths and test the efficiency of our method in measuring the statistically significant length scale of filaments in the mock samples drawn from these simulations . a brief outline of our paper follows .
section 2 describes the data and method of analysis , our results and conclusions are presented in section 3 .
the sloan digital sky survey ( sdss ) @xcite is a wide - field imaging and spectroscopic survey of the sky using a dedicated 2.5 m telescope @xcite with @xmath5 field of view at apache point observatory in southern new mexico .
the survey is carried out in five broad filters namely u , g , r , i and z covering the wavelength range from 3000 to 10000 @xmath6 @xcite .
our present analysis is based on sdss dr6 @xcite galaxy redshift data .
the sdss dr6 includes 9583 @xmath7 imaging and 7425 @xmath7 of spectroscopy .
the spectroscopic data includes @xmath8 spectra with @xmath9 galaxy redshift .
different selection algorithms are used for different categories of sdss targets .
we use the main galaxy sample for the present work .
the main galaxy sample target selection algorithm is detailed in @xcite . the main galaxy sample
comprises of galaxies brighter than a limiting r band petrosian magnitude @xmath10 .
we downloaded the data from catalog archive server ( cas ) of sdss dr6 by structured query language ( sql ) search . for the present analysis we use seven non - overlapping strips each spanning @xmath11 in @xmath12 and @xmath13 in @xmath14 , lying entirely within the survey area of sdss dr6 . @xmath12 and @xmath14
are survey co - ordinates described in @xcite .
these strips are identical in sky coverage as the ones used in @xcite and are shown in figure 1 of that paper .
for each strip we have extracted two different volume limited subsamples in bin1 and bin2 ( table 1 and table 2 ) . for each strip the number of galaxies in bin1 and bin2 are listed in table 1 . the absolute magnitude and redshift limits for bin1 and bin2
are listed in table 2 .
the analysis using `` shuffle '' requires us to cut the entire survey area into squares and shuffle them around .
the thickness of the resulting subsamples increases with redshift . for our analysis
we have considered a smaller region of uniform thickness corresponding to the value at the lowest redshift . for each luminosity
bin the area and number density are listed in table 2 .
for each bin the galaxy number density varies slightly across the seven strips and the average density along with the @xmath15 variation is shown in table 2 .
for all the subsamples , the thickness is much smaller than the other two dimensions and hence it is collapsed along the thickness resulting in a 2d distribution ( figure [ fig:1 ] ) .
we study luminosity dependence of statistically significant length scale of filaments by measuring and comparing it s value in bin1 and bin2 .
we label the galaxies in bin1 as faint and galaxies in bin 2 as bright .
we also separately study the colour and morphology dependence in bin 2 ( table 1 and table 2 ) by dividing the galaxies in this bin into red / blue and early / late type galaxies .
the galaxies in bin 1 are not divided on the basis of their colour or morphology and hence for bin 1 no cut - offs in @xmath16 colour and concentration index @xmath17 are listed in table 2 . [ cols="^,^,^,^,^",options="header " , ] all the strips that we have analyzed are nearly two dimensional .
the strips were all collapsed along the thickness ( the smallest dimension ) to produce 2d galaxy distributions .
we use the 2d `` shapefinder '' statistic @xcite to quantify the average filamentarity of the patterns in the resulting galaxy distribution .
a detailed discussion is presented in paper ii , and we present only the salient features here . the reader is referred to @xcite for a discussion of shapefinders in three dimensions . the galaxy distribution is represented as a set of 1s on a 2-d rectangular grid of spacing @xmath18 , empty cells are assigned a value @xmath19 and filled cells are assigned a value @xmath20 .
we identify connected filled cells using the ` friends - of - friend ' ( fof ) algorithm .
the filamentarity of each cluster is quantified using the shapefinder @xmath21 defined as @xmath22 where @xmath23 and @xmath24 are respectively the perimeter and the area of the cluster , and @xmath25 is the grid spacing .
the shapefinder @xmath21 has values @xmath19 and @xmath20 for a square and filament respectively , and it assumes intermediate values as a square is deformed to a filament .
we use the average filamentarity @xmath26 to asses the overall filamentarity of the clusters in the galaxy distribution .
the distribution of 1s corresponding to the galaxies is sparse .
only @xmath27 of the cells contain galaxies and there are very few filled cells which are interconnected . as a consequence fof fails to identify the large coherent structures which corresponds to filaments in the galaxy distribution .
we overcome this by successively coarse - graining the galaxy distribution . in each iteration of coarse - graining all the empty cells adjacent to a filled cell
( i.e. cells at the 4 sides and 4 corners of a filled cell ) are assigned a value @xmath20 .
this causes clusters to grow , first because of the growth of individual filled cells , and then by the merger of adjacent clusters as they overlap .
coherent structures extending across progressively larger length - scales are identified in consecutive iterations of coarse - graining .
so as not to restrict our analysis to an arbitrarily chosen level of coarse - graining , we study the average filamentarity after each iteration of coarse - graining .
the filling factor @xmath28 quantifies the fraction of cells that are filled and its value increases from @xmath29 and approaches @xmath20 as the coarse - graining proceeds .
we study the average filamentarity @xmath30 as a function of the filling factor @xmath28 ( figure [ fig:2 ] ) as a quantitative measure of the filamentarity at different levels of coarse - graining .
the values of @xmath28 corresponding to a particular level of coarse - graining shows a slight variation from strip to strip . in order to combine and compare the results from different strips , for each strip we have interpolated @xmath30 to @xmath31 values of @xmath28 at an uniform spacing of @xmath32 over the interval @xmath33 to @xmath34 .
coarse - graining beyond @xmath35 washes away the filaments and hence we do not include this range for our analysis .
we use a statistical technique `` shuffle '' to determine the largest length - scale at which the filamentarity is statistically significant .
a grid with squares blocks of side @xmath36 is superposed on the original data slice ( figure [ fig:1 ] ) .
blocks of data which lie entirely within the slice are then randomly interchanged , with rotation , repeatedly , to form a new shuffled slice .
the shuffling process eliminates coherent features in the original data on scales larger than @xmath36 , keeping clustering at scales below @xmath36 nearly identical to the original data .
all the structures spanning length - scales greater than @xmath36 that exist in the shuffled slices are the result of chance alignments . at a fixed value of @xmath36
, the average filamentarity in the original sample will be larger than in the shuffled data only if the actual data has more filaments spanning length - scales larger than @xmath36 , than that expected from chance alignments .
the largest value of @xmath36 , @xmath37 , for which the average filamentarity of the shuffled slices is less than the average filamentarity of the actual data gives us the largest length - scale at which the filamentarity is statistically significant .
filaments spanning length - scales larger than @xmath37 arise purely from chance alignments . for each value of @xmath36 we generated @xmath38 different realization of the shuffled slices .
to ensure that the edges of the blocks which are shuffled around do not cut the actual filamentary pattern at exactly the same place in all the realizations of the shuffled data , we randomly shifted the origin of the grid used to define the blocks .
the values of ff and @xmath30 in the 24 realizations differ from one another and from the actual data at the same stage of coarse - graining .
so as to be able to quantitatively compare the shuffled realizations with the actual data , we interpolate the values of @xmath30 in the shuffled realization at the same values of ff as interpolated for the actual data .
the mean @xmath39 $ ] and the variance @xmath40)^2 $ ] of the average filamentarity was determined for the shuffled data at each value of ff using the 24 realizations .
the difference between the filamentarity of the shuffled data and the actual data was quantified using the reduced @xmath41 per degree of freedom @xmath42-\bar{f_2}[{\rm shuffled}]))_a^2 } { ( \delta f_2[{\rm shuffled}])_a^2}\ ] ] where the sum is over different values of the filling factor ff .
we use `` shuffle '' to determine the statistically significant length scale of filaments in the seven sdss strips in magnitude bin 1 and bin 2 ( table 1 and table 2 ) .
we label bin 1 as faint and bin 2 as bright .
the top two panels of figure [ fig:2 ] shows the results of `` shuffle '' on strip 1 from bin 1 and bin 2 respectively . in top two panels of figure [ fig:2 ]
we see that shuffling the data with @xmath43 causes a large drop in the average filamentarity ( @xmath30 ) .
the average filamentarity increases as the value of @xmath36 increases and the value of average filamentarity of the unshuffled slice comes within @xmath44 error bars when the data is shuffled with @xmath45 and @xmath46 for bright and faint galaxies respectively .
difference between the average filamentarity of shuffled and unshuffled slice is quantified by @xmath47 and the top two panels of figure [ fig:3 ] shows @xmath47 as a function of @xmath36 . in top two panels of figure [ fig:3 ]
we see that @xmath47 approaches 1 at @xmath45 for bright and @xmath46 for faint galaxies .
so in strip 1 from bin1 and bin 2 the filaments are statistically significant upto length scales @xmath1 and @xmath48 respectively .
the results from the other strips are not shown here .
averaging the results from all the nine strips we find that the filaments are statistically significant upto length scales @xmath49 and @xmath50 in bin 1 and bin 2 respectively .
the right and left middle panels of figure [ fig:2 ] shows the results of `` shuffle '' for red and blue galaxies ( table 2 ) from strip 1 in bin 2 .
the right and left middle panels of figure [ fig:2 ] shows that a large drop in the average filamentarity is observed when the data is shuffled with @xmath43 . in
both the panels we see that shuffling the data with @xmath51 increase the average filamentarity and brings it back in agreement with the actual data .
in two middle panels of figure [ fig:3 ] we see that in both class of galaxies @xmath47 approaches 1 at @xmath51 establishing that for both red and blue galaxies the filaments are statistically significant upto length scales @xmath52 . here
we have not shown the results from the other strips . combining results from all the nine strips we find that the filaments are statistically significant upto length scales @xmath53 and @xmath54 for red and blue galaxies respectively .
we show the results of `` shuffle '' for the early and late type galaxies ( table 2 ) in the right and left bottom panels of figure [ fig:2 ] . here
also we see that the average filamentarity decreases when the data is shuffled with @xmath43 and it increases and matches with the actual data when the data is shuffled with @xmath51 .
the @xmath47 approaches 1 at @xmath51 for both early and late type galaxies as shown in two bottom panels of figure [ fig:3 ] .
this establishes that for both early and late type galaxies the filaments are statistically significant upto length scales @xmath55 and not beyond . we have shown here only the results for strip 1 in bin 2 . combining the results from all the strips the filaments
are found to be statistically significant upto length scales @xmath56 and @xmath57 for early and late type galaxies respectively .
it is to be noted in middle and bottom two panels of figure [ fig:2 ] that the red and early type galaxies have higher degree of filamentarity as compared to blue and late type galaxies at smaller values of filling factor .
we reported this effect in @xcite .
it is interesting to note that although the average filamentarity ( @xmath30 ) depends on luminosity , colour and morphology of galaxies , the statistically significant length scale of filaments does not depend on these galaxy properties . in figure [ fig:4 ]
we show the results of `` shuffle '' in simulated sdss ngp strips with different biases .
the top left panel show the result for the dark matter distribution from @xmath0cdm n - body simulations . in this case
galaxies are assumed to exactly trace the dark matter and the bias parameter @xmath58 .
we see that shuffling the simulated data with @xmath43 causes a large drop in the average filamentarity of the simulated galaxy distribution .
with increasing l the value of average filamentarity of the shuffled slice slowly approaches the values corresponding to the unshuflled data .
it is found that at @xmath45 the two are within @xmath59 error bars .
the top right panel shows the result for b=1.2 .
we see a very similar result and the unshuflled and shuffled data agrees well when the data is shuffled with @xmath45 . in top
left and right panels of figure [ fig:5 ] we see that @xmath47 approaches 1 at @xmath45 for both @xmath58 and @xmath60 and hence filaments are statistically significant upto length scales @xmath61 for both cases .
the results are shown for one single mock sdss strip . averaging the results from all the nine strips it is found that for b=1 and b=1.2 filaments are statistically significant upto length scales @xmath62 and @xmath63 respectively .
the bottom left panel of figure [ fig:4 ] shows the result of shuffling on simulated sdss strips for a bias b=0.8 . as usual
shuffling the data with @xmath43 causes a drop in average filamentarity of the shuffled slice but the drop is smaller as compared to other bias values .
as l increases the average filamentarity slowly grows and approaches the values corresponding to the unshuffled data at @xmath45 . in the bottom right panel of figure
[ fig:4 ] results are shown for a high bias b=1.5 .
shuffling shows similar results for b=1.5 but a relatively large drop is seen when the data is shuffled with @xmath43 .
with increasing l the average filamentarity grows relatively faster as compared to b=0.8 and finally the average filamentarity in the shuffled data levels up with the unshuffled data when @xmath51 is used for shuffling . in the bottom left and right
panels of figure [ fig:5 ] it is found that @xmath47 approaches 1 at @xmath45 and @xmath51 indicating that filaments are statistically significant upto length scales @xmath61 and @xmath52 for b=0.8 and b=1.5 respectively .
these are the results shown for a single mock sdss strip and by combining the results from the other strips we get that the filaments are statistically significant upto length scales @xmath64 and @xmath65 for @xmath66 and @xmath67 respectively .
it is to be noted in figure [ fig:4 ] that as the bias @xmath4 is increased the average filamentarity ( @xmath30 ) increases at smaller @xmath28 .
we reported this effect in @xcite .
but interestingly the shuffling length at which the avergae filamentarity of the unshuflled and shuffled data agrees does not depend on bias .
figure [ fig:6 ] shows the statistically significant length scale of filaments as a function of bias .
this figure shows that the statistically significant length scale of filaments is nearly independent of bias .
the mean value is @xmath68 for the bias values @xmath69 .
it tends to decrease for a high value of bias @xmath70 .
this could be possibly related to the fact that at very high values of bias we preferentially identify only very high density regions which are mostly related to cluster like structures rather than filaments .
earlier in paper i we established that the filaments are statistically significant upto length scales of @xmath61 .
this measured value lies well within @xmath15 error bars of the measured values for mock galaxy distribution from @xmath0cdm n - body simulations .
we conclude that the @xmath0cdm is consistent with sdss observations .
the galaxy samples constructed over different magnitude ranges have different area and galaxy number density ( table 2 ) .
earlier analysis in paper ii shows that the average filamentarity of the galaxy distribution depends on the area and galaxy number density of the galaxy samples and filamentarity between different galaxy samples can only be compared when the samples have same area , geometry and number density of galaxies .
so it is important to know if the statistically significant length scale of filaments depends on these factors before we can compare it s measured values among various galaxy samples having different areas and galaxy number densities .
we prepare two different set of mock galaxy samples namely area 1 and area 2 ( table 3 ) which have the same galaxy number density as the sdss ngp strip ( area 3 ) but cover different redshift ranges and hence have different areas as listed in table 3 .
we extract three mock galaxy samples from each of the @xmath0cdm dark matter simulation for each set of galaxy samples area 1 and area 2 .
we determine the statistically significant length scale of filaments for these galaxy samples .
the analysis for the mock sdss ngp strip which we name as area 3 in table 3 are already done .
we plot the statistically significant length scale of filaments as a function of the area of the galaxy samples in figure [ fig:7 ] . from this figure
we see that the samples of smaller area tend to have a smaller length scale of filaments . but this trend is weak and the size of the error bars are quite large . we conclude that the statistically significant length scale of filaments are nearly independent of the area of the samples .
we next prepare two different set of mock galaxy samples namely density 1 and density 3 ( table 3 ) which have the same area as the sdss ngp strip ( density 2 ) but have different number densities listed in table 3 .
we extract three mock galaxy samples from each of the @xmath0cdm dark matter simulation for each set of galaxy samples density 1 and density 3 .
we determine the statistically significant length scale of filaments for these galaxy samples .
the analysis for the mock sdss ngp strip which we name as density 2 in table 3 are already done .
we plot the statistically significant length scale of filaments as a function of the number density of galaxies of the samples in figure [ fig:8 ] .
we see a trend of low density samples to have a smaller length scale of filaments .
but the error bars are quite large and the dependence is weak . we conclude that the statistically significant length scale of filaments weakly depends on the number density of galaxies . in a recent work with sdss lrg samples @xcite having much larger area and lower density we find that the statistically significant length scale of filaments does not change much as compared to it s value measured in sdss main galaxy samples .
thus the statistically significant length scale of filaments is a more robust statistics than the average filamentarity .
it does not depend on properties ( size and number density ) of galaxy samples , and physical properties of galaxies .
it emerges as a robust measure of the large scale structures .
the statistically significant length scale of filaments could be also thought of as an indicator of the scale beyond which the galaxy distribution become homogeneous because it tells us the length scale beyond which no coherent structures exist in the galaxy distribution .
multifractal analysis in sdss dr1 @xcite and recently in sdss dr6 @xcite show that the galaxy distribution becomes homogeneous at a length scale between @xmath71 to @xmath72 .
@xcite considered the transition to homogeneity in @xmath0cdm simulations with different bias @xmath73 and find that irrespective of the bias values the distribution become homogeneous at a length scale between @xmath71 to @xmath74 .
these results are consistent with our findings .
finally we test the reliability of our method by applying it to different controlled samples ( table 4 ) of 3d segment cox process .
we use segments of two different length @xmath75 and @xmath76 for simulating the 3d segment cox process .
two different sets of 2d mock samples drawn from these 3d simulations ( table 4 ) are separately analyzed .
the results are shown in top two panels of figure [ fig:9 ] and figure [ fig:10 ] .
we see in top left panel of figure [ fig:9 ] that filamentarity decrease by a very small amount when the mock samples having @xmath75 are shuffled with @xmath43 and there are virtually no drop in filamentarity when the data is shuffled with @xmath77 .
the top right panel of the same figure shows the result of `` shuffle '' on cotrolled samples of segment cox process with segment length @xmath78 .
we see a significant drop in the average filamentarity when the data is shuffled with @xmath43 and the average filamentarity finally saturates to the values corresponding to the unshuffled data when @xmath79 is used for shuffling .
the @xmath47 as a function of @xmath36 for these two cases are shown in top two panels of figure [ fig:10 ] .
we see that @xmath47 approaches 1 at @xmath77 and @xmath79 in these two cases respectively .
this indicates that for these 2d mock samples having @xmath75 and @xmath78 , the filaments are found to be statistically significant upto length scales @xmath80 and @xmath81 respectively . here
the results are shown only for a single simulated mock sample . by averaging the results from all the nine simulated mock samples we find that in the above two cases the filaments are statistically significant upto length scales @xmath82 and @xmath83 respectively .
we note that the values of statistically significant length scale of filaments recovered by our method for straight segments simulated here are a little shorter but close to the original values of @xmath25 used in the simulations .
the simulations are carried out in 3d boxes and our mock samples are the 2d projections of nearly two dimensional samples drawn from these boxes .
this possibly destroys and shortens a lot of segments and the effects are expected to be more prominent for longer segments as reflected in our results .
this effect indicates that also for the 2d real galaxy samples from sdss which are constructed out of a 3d galaxy distribution , shuffle possibly underestimates the values of statistically significant length scale of the real 3d filaments .
the effects of slicing and projections can be avoided if we simulate the segment cox process in 2d as in this case the segments lie on the same plane ( figure [ fig:11 ] ) .
we apply shuffle to a set of mock 2d strips constructed out of simulations of 2d segment cox process ( table 5 ) .
we use two different segment lengths @xmath84 and @xmath76 to simulate the 2d segment cox process .
we use the same segment lengths as used in the simulations of 3d segment cox process so as to test the effect of slicing and projection .
we show the results in bottom two panels of figure [ fig:9 ] and figure [ fig:10 ] .
bottom left panel of figure [ fig:9 ] shows that when the mock samples with @xmath84 are shuffled with @xmath43 the average filamentarity of the shuffled data falls below that of the unshuffled data .
the average filamentarity of the shuffled data saturates to the unshuffled data when the data is shuffled with @xmath85 .
the bottom right panel of figure [ fig:9 ] shows a significant drop in the average filamentarity when the mock samples with @xmath76 are shuffled with @xmath43 .
the average filamentarity of the shuffled data increases slowly with increasing shuffling length and finally saturates to the unshuffled data at shuffling length @xmath45 .
we see in bottom two panels of figure [ fig:10 ] the @xmath47 as a function of @xmath36 approaches 1 at @xmath85 and @xmath45 in these two cases respectively .
this indicates that for these controlled samples of 2d segment cox process having @xmath75 and @xmath86 , the filaments are statistically significant upto length scales @xmath87 and @xmath61 respectively .
we have shown the results for a single simulated strip in these plots .
we combine the results from all the nine simulated strips and find that the filaments are statistically significant upto length scales @xmath88 and @xmath89 for simulations of 2d segment cox process with segment length @xmath75 and @xmath76 respectively .
we note that our method can successfully recover better the length of inputted segments when the mock samples are drawn from simulations of 2d segment cox process where the effects of slicing and projections are absent
. there would be another effect due to the boundaries of the samples which again shortens the length of some segments near the boundaries but this effect is not so important in our analysis as the extent of the samples are much larger than the length of the inputted segments .
this analysis suggets that the actual length scale of the real 3d filaments can be recovered with shuffle if the analysis is extended in 3d .
we propose to take up this in a future work . in conclusion
we note that although the average filamentarity of the galaxy distribution depends on different galaxy properties , the statistically significant length scale of filaments does not depend on luminosity , colour and morphology of galaxies .
we find that the measured values of the statistically significant length scale of filaments in sdss are consistent with that measured from @xmath0cdm simulations .
we considered the @xmath0cdm model with different values of bias and find that the statistically significant length scale of filaments is nearly independent of bias .
different class of galaxies are differently biased with respect to underlying dark matter distribution and this result is possibly related to the fact that statistically significant length scale of filaments is nearly the same for different class of galaxies .
unlike the average filamentarity , the statistically significant length scale of filaments is nearly independent of the area and galaxy number density of the samples .
this establishes the statistically significant length scale of filaments as a robust statistics of galaxy distribution .
bp acknowledges somnath bharadwaj for useful discussions and suggestions . thanks are due to an anonymous referee , whose comments and suggestions led to significant improvement of the paper .
the sdss dr6 data was downloaded from the sdss skyserver http://cas.sdss.org/dr6/en/ .
funding for the creation and distribution of the sdss archive has been provided by the alfred p. sloan foundation , the participating institutions , the national aeronautics and space administration , the national science foundation , the u.s .
department of energy , the japanese monbukagakusho , and the max planck society .
the sdss web site is http://www.sdss.org/. the sdss is managed by the astrophysical research consortium ( arc ) for the participating institutions .
the participating institutions are the university of chicago , fermilab , the institute for advanced study , the japan participation group , the johns hopkins university , the korean scientist group , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , university of pittsburgh , princeton university , the united states naval observatory , and the university of washington . | we have used a statistical technique shuffle " @xcite in seven nearly two dimensional strips from the sloan digital sky survey data release six ( sdss dr6 ) to test if the statistically significant length scale of filaments depends on luminosity , colour and morphology of galaxies .
we find that although the average filamentarity depends on these galaxy properties , the statistically significant length scale of filaments does not depend on them .
we compare it s measured values in sdss against the predictions of @xmath0cdm n - body simulations and find that @xmath0cdm model is consistent with observations .
the average filamentarity is known to be very sensitive to the bias parameter . using @xmath0cdm n - body simulations
we simulate mock galaxy distributions for sdss ngp equatorial strip for different biases and test if the statistically significant length scale of filaments depends on bias .
we find that statistically significant length scale of filaments is nearly independent of bias .
this result is possibly related to the fact that statistically significant length scale of filaments is nearly the same for different class of galaxies which are differently biased with respect to underlying dark matter distribution .
the average filamentarity is also known to be dependent on the galaxy number density and size of the samples .
we use @xmath0cdm dark matter n - body simulations to test if the statistically significant length scale of filaments depends on number density of galaxies and size of the samples .
our analysis shows that the statistically significant length scale of filaments very weakly depends on these factors .
finally we test the reliability of our method by applying it to controlled samples of segment cox process and find that our method successfully recovers the length of the inputted segments . summarizing these results we conclude that the statistically significant length scale of filaments is a robust measure of the galaxy distribution . methods : numerical - galaxies : statistics - cosmology : theory - cosmology : large scale structure of universe |
among the essential goals of qcd spin physics one has first , to understand the nucleon spin structure in terms of its basic partonic constituents and second , to test the _
spin sector _ of perturbative qcd , at the highest possible precision level . concerning the first point
, one needs to know how the quark and gluon distributions in a polarized nucleon make its spin one - half and several questions arise in particular : what is the role of the orbital angular momentum ?
the second point is very relevant to reinforce the validity of the already well established perturbative qcd theory , because many spin asymmetries have been calculated , at the next - to - leading order ( nlo ) , and have not yet been compared with experimental data .
therefore it is very legitimate to ask to what extent they will agree .
we will try to answer the following questions : what is known ? what is missing ?
what needs to be measured next ?
what are the prospects ?
+ the basic information comes from deep inelastic scattering ( dis ) , unpolarized @xmath0 , or polarized @xmath1 . in the unpolarized case , widely measured over the last three decades , one gets access to @xmath2 $ ] . here
the @xmath3 s ( same for antiquarks ) are defined as @xmath4 , where @xmath5 are the quark distributions in a polarized proton with helicity parallel @xmath6 or antiparallel @xmath7 to that of the proton . in the polarized case ,
one measures the corresponding polarized structure function , @xmath8 $ ] .
similarly @xmath9 s ( same for antiquarks ) are defined as @xmath10 .
the gluon distributions are also defined as @xmath11 and @xmath12 , but in dis they are not accessible directly and only enter in the qcd @xmath13 evolution of the quark distributions . + there is a long list of interesting topics , _
e.g. _ characteristic features of unpolarized and polarized parton distributions , flavor separation of @xmath14 , @xmath15 , gluon polarization in the nucleon , generalized parton distributions , quark transversity @xmath16 and double transverse spin asymmetries @xmath17 , single spin asymmetries ( ssa ) @xmath18 and qcd mechanisms , etc ... + in this opening lecture , for lack of time , we will have to make a strong selection , but given the high density of the scientific program , it will certainly allow to cover all missing important subjects .
a new set of parton distribution functions ( pdf ) was constructed in the framework of a statistical approach of the nucleon @xcite , which has the following characteristic features : + - for quarks ( antiquarks ) , the building blocks are the helicity dependent distributions @xmath5 ( @xmath19 ) and we define @xmath20 and @xmath21 ( similarly for antiquarks ) .
+ - at the initial energy scale taken at @xmath22 , these distributions are given by the sum of two terms , a quasi fermi - dirac function and a helicity independent diffractive contribution , which leads to a universal behavior for all flavors at very low @xmath23 .
+ - the flavor asymmetry for the light sea , _ i.e. _ @xmath24 , observed in the data is built in . this
is clearly understood in terms of the pauli exclusion principle , based on the fact that the proton contains two @xmath25 quarks and only one @xmath26 quark .
+ - the chiral properties of qcd lead to strong relations between @xmath27 and @xmath28 .
+ for example , it is found that the well estalished result @xmath29 implies @xmath30 and similarly @xmath31 leads to @xmath32 . + - concerning the gluon , the unpolarized gluon distribution is given in terms of a quasi bose - einstein function , with _ no free parameter _ , but for simplicity , one assumes zero gluon polarization , _ i.e. _ @xmath33 , at the initial energy scale .
+ - all unpolarized and polarized distributions depend upon _ eight _ free parameters , which were determined in 2002 ( see @xcite ) , from an nlo fit of a selected set of accurate dis data .
+ for illustration , the @xmath34 light quark ( antiquark ) distributions are displayed on fig . 1 and we clearly notice the essential features mentioned above .
more recently , new tests against experimental ( unpolarized and polarized ) data turned out to be very satisfactory , in particular in hadronic reactions @xcite . + on the left ( right ) the light quark ( antiquark ) distributions with different helicities versus @xmath23 for @xmath35 , taken from ref .
@xcite.,title="fig:",width=245 ] on the left ( right ) the light quark ( antiquark ) distributions with different helicities versus @xmath23 for @xmath35 , taken from ref .
@xcite.,title="fig:",width=245 ] the statistical approach has been extended to the interesting case where the pdf have , in addition to the usual bjorken @xmath23 dependence , an explicit transverse momentum @xmath36 dependence @xcite and this might be used in future calculations with no @xmath36 integration . concerning the strange quark and antiquark distributions , a simplifying assumption consists to take @xmath37 and similarly for the corresponding polarized distributions @xmath38 .
however a careful analysis of the data led us to the conclusion that @xmath39 and the corresponding polarized distributions are unequal , small and negative @xcite . + now let us come back to the important prediction of the statistical approach , namely @xmath40 and @xmath32 , which contrasts with the flavor symmetric assumption @xmath41 made , for example , in ref .
@xcite . with this assumption ,
the @xmath15 do nt contribute to the bjorken sum rule , so one has to increase the absolute values of the valence contributions to @xmath42 and @xmath43 , in order to satisfy this sum rule .
as shown on fig . 2 , this leads to over estimate @xmath44 in the valence region , but it is not the case for the statistical approach .
this has been confirmed by recent compass data @xcite .
+ various recent data on the isovector structure function @xmath44 compared to the statistical model prediction ref . @xcite ( lower curve ) and the aac calculation ref .
@xcite ( upper curve)(taken from ref .
@xcite).,width=302 ] left : quark and antiquark polarized parton distributions as a function of @xmath23 for @xmath45 .
data from ref .
right : ratios @xmath46 and @xmath47 as a function of @xmath23 . data from hermes for @xmath45 @xcite and a jlab experiment @xcite .
in both the curves are predictions from the statistical approach ref .
@xcite.,title="fig:",width=245 ] left : quark and antiquark polarized parton distributions as a function of @xmath23 for @xmath45 .
data from ref .
right : ratios @xmath46 and @xmath47 as a function of @xmath23 .
data from hermes for @xmath45 @xcite and a jlab experiment @xcite . in both
the curves are predictions from the statistical approach ref .
@xcite.,title="fig:",width=245 ] there is another way to test directly the predictions of the statistical approach for the polarized quark distributions and their flavor separation .
this has been obtained from the semi - inclusive polarized dis and the hermes data are shown on the left hand side of fig .
3 . on the right hand side of fig . 3
, we also display the very accurate jlab data which show that , even in the high @xmath23 region , @xmath48 remains positive whereas @xmath49 remains negative , in accordance with the statistical approach expectations .
+ these features can and will be also investigated in future runs with polarized @xmath50 collisions at bnl - rhic , which we briefly discuss now . + consider the parity - violating helicity asymmetry @xmath51 @xmath52 where @xmath34 stands for the helicity of one polarized proton beam and @xmath53 is the @xmath54 rapidity . for @xmath55 , at the lowest order of the drell - yan production mechanism , it reads @xmath56 where @xmath57 , @xmath58 and @xmath59 . for @xmath60 production one interchanges @xmath25 and @xmath26 .
the general trend of @xmath61 can be easily understood and , for example at @xmath62 near @xmath63 , @xmath64 and @xmath65 , evaluated at @xmath66 .
similarly for near @xmath67 , @xmath68 and @xmath69 , evaluated at @xmath70 .
+ the features appear clearly on the left hand side of fig . 4 , where the calculations were done at two different energies . for completeness
we also show the predicted @xmath71 on the right hand side of fig . 4 , but in this case the interpretation is not so straightforward .
moreover the production rate of @xmath72 s is much lower than @xmath54 s .
+ left : predictions from the statistical appoach for the parity violating asymmetry @xmath73 for the @xmath74 production , versus the @xmath54 rapidity @xmath53 , at @xmath75gev ( dashed curve ) and @xmath76gev ( solid curve ) .
right : same for the @xmath77 production .
( taken from ref .
@xcite),title="fig:",width=245 ] left : predictions from the statistical appoach for the parity violating asymmetry @xmath73 for the @xmath74 production , versus the @xmath54 rapidity @xmath53 , at @xmath75gev ( dashed curve ) and @xmath76gev ( solid curve ) .
right : same for the @xmath77 production .
( taken from ref .
@xcite),title="fig:",width=245 ] however there is an important point to mention here , since the @xmath54 s are not directly seen . for the most relevant signature , if one selects the leptonic decay @xmath78 , one measures in fact @xmath79 where @xmath80 is the charged lepton rapidity .
fortunately , by using the rhicbos code due to p. nadolski , one finds that @xmath81 has essentially the same trend as @xmath61 .
+ so much for the quarks , let us now turn to the gluon distributions and we first consider the unpolarized distribution @xmath82 . in the statistical approach it has a very simple expression ( see ref .
@xcite ) , which is consistent with the available data , most coming indirectly from the qcd @xmath13 evolution of @xmath83 , defined earlier , in particular in the low @xmath23 region
. however it is known that @xmath84 dis cross section is characterised by two independent structure funtions , @xmath83 and the longitudinal structure function @xmath85 . for low @xmath13 ,
the contribution of the later to the cross section at hera is only sizeable at @xmath23 smaller than approximately @xmath86 and in this domain the gluon density dominates over the sea quark density .
more precisely , it was shown that using some approximations , one has @xcite
@xmath87 \simeq \frac{8.3}{\alpha_s}f_l(0.4x , q^2)~.\ ] ] before hera was shut down , a dedicated run period with reduced proton beam energy was approved and we are waiting for these new h1 results on @xmath88 .
we show on fig .
5 the predictions of the statistical approach and the new data , whose precision is expected to be rather good , will allow to test its predictive power , once more .
+ statistical approach predictions for the longitudinal structure function @xmath85 with earlier h1 data ( taken from ref .
@xcite).,width=302 ] the polarized gluon distribution @xmath89 is also extremely important to determine and we have the following helicity sum rule @xmath90 where @xmath91dx$ ] is twice the quark ( + antiquark ) contribution to the nucleon helicity and @xmath89 , @xmath92 are the contributions of gluon and orbital angular momentum of quark and gluon . so far @xmath93 and the sum rule is not satisfied .
+ there are several attempts to extract @xmath94 from dis using different processes and the most recent results will be presented later in this workshop .
the rhic spin program is also putting a high priority to this determination and the cleanest reaction is inclusive prompt photon production , which is dominated by the subprocess @xmath95 . the double helicity asymmetry , which has schematically the following expression @xmath96}{\sum_q e_q^2 \left[q(x_2)+ \bar{q}(x_2 ) \right ] } \bigg ] \cdot \hat{a}_{ll}(gq \rightarrow \gamma q ) + ( 1\leftrightarrow 2)~,\ ] ] is directly proportional to @xmath89 .
this has not been measured yet , but from the measurements on @xmath97 , we have all indications that @xmath89 is small and still badly known , unfortunately .
+ the next very serious question is indeed : are there relevant contributions from @xmath92 ?
the existence of this new quark distribution @xmath16 , was first mentioned by ralston and soper in 1979 , by studying the angular distribution in @xmath98 with transversely polarized protons .
it was merely forgotten until 1990 , where it was first realized that it completes the description of the quark distribution in a nucleon as a density matrix @xmath99 this quark transversity @xmath16 is chiral odd , leading twist and decouples from dis .
so it was never measured and we only have the following positivity bound @xcite @xmath100 which survives up to nlo corrections .
it is indeed accessible in @xmath98 , with both protons transversely polarized .
the double transverse spin asymmetry @xmath17 reads @xmath101 where @xmath102 and @xmath103 is the dilepton mass square .
it involves the product of @xmath104 and @xmath105 , as expected from the dominant @xmath106 annihilation drell - yan mechanism .
predictions using the saturation of the bound , lead to some estimates of only a few percents , but it is on the list of future measurements at the bnl - rhic spin program . + the asymmetry at the @xmath72 pole , which reads @xmath107 is also expected to be small .
however , for the @xmath108 production , considered above , @xmath109 , because the @xmath54 has a @xmath110 coupling , _ i.e. _ @xmath111 , which remains to be checked .
+ there is no such a transversity distribution for gluons which carry a spin one and this fact has important consequences for @xmath17 of different reactions .
for example in the case of single - jet production , according to pqcd , the cross section in the low @xmath112 region is dominated by gluon- gluon collisions , in the medium @xmath112 region by gluon - quark collisions and in the high @xmath112 region by quark - quark collisions . as a result , @xmath17 is expected to be non - zero only in this last kinematic region and this is what we see on the left hand side of fig .
we have a similar situation for prompt photon production , shown on the right hand side of fig .
6 . these results , which were obtained by using the positivity bound , probe the sensitivity only to quark transversity in the hight @xmath112 region . as was noticed in ref .
@xcite , we expect double spin transverse asymmetries to be much smaller than double helicity asymmetries , _
i.e. _ @xmath113 and this theoretical observation must be carefully confirmed experimentally . left : upper bounds for @xmath17 for single jet production at rhic , with the expected statistical errors .
right : same for prompt photon production ( taken from ref .
@xcite).,title="fig:",width=298 ] left : upper bounds for @xmath17 for single jet production at rhic , with the expected statistical errors .
right : same for prompt photon production ( taken from ref .
what is a single spin asymmetry ( ssa ) ? + consider the collision of a proton of momentum @xmath114 , carrying a transverse spin @xmath115 and producing an outgoing hadron with transverse momentum @xmath116 .
the ssa defined as @xmath117 is zero , unless the cross section contains a term @xmath118 .
it can be shown that this requires the existence of an _ helicity flip _ and _ final state interactions _ , which generate a phase difference between the flip and the non - flip amplitudes , to avoid violation of time reversal invariance . in the naive parton model one
expects very small ssa , because of the double suppression @xmath119 , where @xmath120 is the quark mass and @xmath121 the energy scale of the process .
+ actually a large ssa has been discovered 30 years ago at fnal with a 300 gev / c unpolarized proton beam in @xmath122 @xcite and many
more ssa have been observed later , in particular large ssa in @xmath123 and @xmath124 at fnal by e704 @xcite and more recently by star at bnl - rhic @xcite .
these data have the same trend , as shown in fig . 7 , although they were obtained in very different energy ranges
. therefore one can be tempted to conclude that they originate from the same mechanism satisfying scaling . +
before discussing this point , we recall that in the collinear approximation , the mechanism to generata ssa is based on higher - twist quark - gluon correlators ( efremov - teryaev 1982 , qiu - sterman 1991 ) .
however , if one introduces transverse momentum dependence ( tmd ) , two qcd mechanisms have been proposed : + - tmd parton distributions @xmath125 sivers effect 1990 + - tmd fragmentation distributions @xmath125 collins effect 1993 + the gauge - invariance properties of the tmd pdf have been first clarified for dis and drell - yan processes in ref . @xcite . in general
both sivers and collins effects contribute to a specific reaction , although there are some cases in which only one of them contributes . for example in semi inclusive dis , the collins effect is the only mechanism that can lead to asymmetries @xmath126 and @xmath127 . on the other hand , it does not appear in some electroweak interaction processes , where there is only the sivers effect . in prompt photon production in @xmath50 collisions , which is dominated by @xmath128 ,
the ssa is sensitive to either the quark or the gluon sivers functions , according to the value of the photon @xmath129 @xcite . + now let us ask : do we understand the ssa displayed on fig . 7 , given the fact that star is at a very small angle 2.6 deg . , whereas e704 is at a much larger angle , between 9 deg . and 64 deg . ?
a negative answer is partially obtained by looking at the cross section .
the pqcd nlo calculation underestimates the cross section at _ low _ energies and _ medium _ angles , namely for the e704 kinematic region .
this is shown on fig .
7 and it means that one should not ignore other contributions .
this is not the case at 90 deg . and at very small angles at high energy , which is the star kinematic range .
to conclude , one should not try to `` explain '' the ssa , ignoring the unpolarized cross section @xcite .
of course one should not forget resummation effects , which might help clarifying the situation .
left : the single spin asymmetry @xmath18 as a function of @xmath129 at two different energies .
the data are from refs .
right : a comparison between a pqcd nlo calculation and data for two different angles ( taken from ref .
@xcite ) .,title="fig:",width=264 ] left : the single spin asymmetry @xmath18 as a function of @xmath129 at two different energies .
the data are from refs .
right : a comparison between a pqcd nlo calculation and data for two different angles ( taken from ref .
@xcite ) .,title="fig:",width=264 ] + this work was completed during my visit at the universidad santa maria , valparaiso , supported by the cooperation program ecos - conicyt c04e04 between france and chile .
+ i am grateful to the organizers of dspin07 , for their invitation to this conference dedicated to l. i. lapidus , i had the great privilege to meet several times .
my special thanks go also to prof .
efremov for providing a full financial support and for making , once more , this meeting so successful .
+ 99 c. bourrely , f. buccella and j. soffer , euro .
j. * c23 * , 487 ( 2002 ) .
c. bourrely , f. buccella and j. soffer , modern phys .
a18 * , 771 ( 2003 ) .
c. bourrely , f. buccella and j. soffer , euro .
j. * c41 * , 327 ( 2005 ) .
c. bourrely , f. buccella and j. soffer , modern phys .
a21 * , 143 ( 2006 ) .
c. bourrely , f. buccella and j. soffer , phys .
* b648 * , 39 ( 2007 ) . h. hirai , s. kumano and n. saito , phys .
rev . * d74 * , 014015 ( 2006 ) .
c. bourrely , f. buccella and j. soffer , ( in preparation ) .
compass coll .
, m. alekseev et _ al .
_ , arxiv:0707.4077 [ hep - ex ] .
hermes coll .
, k. ackerstaff et _ al .
_ , phys . lett . *
b464 * , 123 ( 1999 ) .
jlab hall a coll .
, x. zheng , et _ al .
_ , phys . rev . *
c70 * , 065207 ( 2004 ) .
hermes coll .
, a. airapetian et _ al .
lett . * 92 * , 012005 ( 2004 ) .
a. m. cooper - sarkar et _ al .
_ , z. phys . *
c39 * , 281 ( 1988 ) .
j. soffer , phys .
* 74 * , 1292 ( 1995 ) .
x. artru , m. elchikh , j. m. richard , j. soffer and o. teryaev , _ spin observables and spin structure functions : inequalities and dynamics _ , physics reports ( 2008 )
. j. soffer , m. stratmann and w. vogelsang , phys .
rev . * d65 * , 114024 ( 2002 ) .
a. mukherjee , m. stratmann and w. vogelsang , phys .
rev . * d67 * , 114006 ( 2003 ) .
g.bunce et al . , phys . rev .
* 36 * , 1113 ( 1976 ) .
e704 coll .
, d. l. adams , phys
b261 * , 201 ( 1991 ) .
star coll .
, j. adams et _ al .
* 92 * , 171801 ( 2004 ) .
s. brodsky , d.s .
hwang and i. schmidt , phys .
b530 * , 99 ( 2002 ) .
i. schmidt , j. soffer and j. j. yang , phys .
b612 * , 258 ( 2005 ) .
c. bourrely and j. soffer , euro .
j. * c36 * , 371 ( 2004 ) . | our knowledge of the nucleon spin structure has greatly improved over the last twenty years or so , but still many fundamental questions remain unsolved .
i will try to review some of the puzzling aspects of the structure of the nucleon spin , in particular , what is known , what remains to be discovered and the prospects for the near future .
i will also focus on some current activities in qcd spin physics . *
the nucleon spin structure and qcd spin physics jacques soffer * |
the meson production reaction in nucleon - nucleon collisions near threshold has the potential to gain new information on hadron properties @xcite , and the experimental database on meson production in nucleon - nucleon collisions has expanded significantly in recent years . on the other hand ,
the study of the strangeness content of the quark wave functions of baryons and baryon resonances , not only in experimental side but also in theoretical side , has been an interesting area @xcite , which is expected to provide new information on the configuration of baryons and baryon resonances . in the naive quark model ,
the nucleon and nucleon resonances have no strangeness contents , whereas the @xmath0 meson is an ideally mixed pure @xmath12 state . from the point of view of the naive quark model
the @xmath13 reaction involves disconnected quark lines and is an okubo - zweig - iizuka ( ozi ) rule @xcite suppressed process .
the study of @xmath0 meson production in nucleon - nucleon reactions may provide information on the strangeness degrees of freedom in the nucleon or nucleon resonances and is of importance both experimentally and theoretically .
several years ago , the exclusive production cross section for @xmath0 meson production in @xmath14 collisions at p@xmath15 = 3.67 gev / c was measured by the disto collaboration @xcite , and the preliminary result at an excess energy of 18.5 mev above the threshold was also published by the anke group @xcite . with these experimental information about this reaction , several theoretical papers @xcite by using various models were published to try to explain the experimental data .
recently , more data at other energies are available from the anke facility @xcite . comparing the data for the @xmath16 meson production from literature , a significant enhancement of @xmath17 ratio of a factor 8
is found compared to predictions based on the ozi rule .
this findings require more theoretical work to understand its origin .
it is well - known that the @xmath3 resonance couples strongly to the @xmath5n channel .
recently , it was found that the @xmath3 resonance has a significant coupling to @xmath18 in the analysis of the @xmath19 decay and the @xmath20 reaction near threshold @xcite .
the analyses @xcite of the recent saphir and clas @xmath21 data @xcite also show a large coupling of the @xmath3 to @xmath22 . in a chiral unitary coupled channel approach
it was found that the @xmath3 resonance is dynamically generated as a pole in the second riemann sheet with its mass , width , and branching ratios in fair agreement with experiments and the couplings of the @xmath3 resonance to @xmath23 , @xmath24 and @xmath18 are large compared to the @xmath25 channel @xcite .
the analyses of data on the @xmath26 photo - production on the proton for photon energies from 1.527 to 2.227 gev also suggest the coupling of the @xmath27 channel to the @xmath3 resonance @xcite . from the naive quark model ,
both @xmath5 meson and @xmath26 meson have a @xmath12 component .
it seems that the @xmath3 couples strongly to mesons with strangeness or with @xmath12 components .
these phenomena indicate that there may be a significant @xmath12 configuration in the quark wave function of the @xmath3 resonance .
so , we expect that the @xmath3 resonance may also have a significant coupling to the @xmath28 channel . in this paper
, we assume that the productions of the @xmath0 meson in proton - proton and @xmath1 collisions are predominantly through the excitation and decay of the sub-@xmath28-threshold @xmath3 resonance . by using this picture ,
we calculate the @xmath29 and @xmath10 reactions in the framework of an effective lagrangian approach . by comparing with the experimental data we find that the coupling of the @xmath3 resonance to the @xmath28 channel needs to be somewhat larger than its the coupling to @xmath30 channel .
the significant coupling of the @xmath3 resonance to @xmath11 is compatible with previous indications of a large @xmath12 component in the quark wave function of the @xmath3 resonance and may be the real origin of the significant enhancement of the @xmath0 production over the naive ozi - rule predictions . in the next section , we will give the formalism and ingredients in our calculation , then numerical results and discussions are given in sect.3 .
a short summary is given in the last section .
we study the @xmath7 and @xmath10 reactions near threshold in an effective lagrangian approach .
we assume that the near threshold @xmath0 productions in proton - proton and @xmath31 collisions are through the intermediate excitation of the sub-@xmath28-threshold @xmath3 resonance .
the @xmath4 , @xmath5 and @xmath6-meson exchanges are considered for proton - proton collisions . the basic feynman diagrams for the @xmath7 reaction and the s - channel diagram for the @xmath10 reaction are depicted in fig . [ diagram ] and fig . [ pipdiagram ] , respectively . reaction .
the diagram on the left shows the direct process , while that on the right shows the exchange one .
@xmath32 ( @xmath33=1,2,3,4,5 ) stands for the 4-momentum of the initial and final particle ; @xmath34 and @xmath35 stand for the 4-momentum of exchange meson and the intermediate resonance ( @xmath3 ) , respectively.,width=336,height=240 ] reaction .
@xmath36 and @xmath35 stand for the 4-momentum of @xmath37 , proton , @xmath0 , neutron and intermediate resonance ( @xmath3 ) , respectively.,width=240,height=144 ] we use the commonly used interaction lagrangians for @xmath38 , @xmath39 and @xmath40 couplings , @xmath41 @xmath42 @xmath43 at each vertex a relevant off - shell form factor is used . in our computation
, we take the same form factors as that used in the well - known bonn potential model @xcite @xmath44 with n=1 for @xmath4 and @xmath5-meson ; n=2 for @xmath6-meson .
@xmath45 , @xmath46 and @xmath47 are the 4-momentum , mass and cut - off parameters for the exchanged - meson ( @xmath48 ) , respectively . the coupling constants and the cutoff parameters are taken as @xcite : @xmath49 = 14.4 , , @xmath50 = 0.9 , @xmath51 = @xmath52 = 1.3 gev , @xmath53 = 1.6 gev , and @xmath54 = 6.1 .
the value of @xmath39 coupling constant is extremely uncertain , with values of @xmath55 between 0 and 7 being quoted in the literature , we use @xmath55 = 0.4 since many authors say that it is small ( see e.g. @xcite and @xcite ) . to calculate the invariant amplitudes of diagrams in the fig .
[ diagram ] and fig .
[ pipdiagram ] with the @xmath3 resonance model , we also need to know the interaction lagrangians involving the @xmath3 resonance . in ref .
@xcite , a lorentz covariant orbital - spin ( l - s ) scheme for @xmath56 couplings has been illustrated in detail . with this scheme , we can easily write the effective @xmath57 , @xmath58 , @xmath59 and @xmath60 couplings , @xmath61 @xmath62 @xmath63 @xmath64 here @xmath65 and @xmath66 are the rarita - schwinger spin wave functions for the nucleon and @xmath3 resonance ; @xmath67 and @xmath68 are the polarization vectors of the @xmath69 and @xmath0-meson , respectively .
it is worth noting that since the spins of the @xmath69 meson and @xmath0 meson are 1 , both s - wave and d - wave l - s couplings are possible for the @xmath59 and @xmath70 interactions .
it was found that the s - wave coupling has a significant contribution to the partial decay width of the @xmath3 resonance compared with the d - wave @xcite . in our calculation
we only consider the s - wave @xmath3 resonance couplings to @xmath71 and neglect the d - wave @xmath3 resonance couplings to @xmath11 for simplicity .
the monopole form factors for @xmath3-@xmath72-meson vertexes are used , @xmath73 with @xmath74 = @xmath75 = @xmath76 = 1.3 gev .
the @xmath77 , @xmath78 and @xmath79 coupling constants are determined from the experimentally observed partial decay widths of the @xmath3 resonance . with the effective interaction lagrangians described by eq .
( [ pion1 ] ) and eq .
( [ eta1 ] ) , the partial decay widths @xmath80 and @xmath81 can be easily calculated @xcite .
the coupling constants are related to the partial decay widths , @xmath82 where @xmath83 and @xmath84 for the @xmath59 coupling constant , we get it from the partial decay width @xmath85 , and the partial decay width can be evaluated from the total invariant amplitude @xmath86 of the @xmath87 decay and a three - body phase space integration , @xmath88 where @xmath89 is the propagator of the @xmath69 meson with the form @xmath90 here @xmath35 and @xmath91 are the 4-momentum of the @xmath3 resonance and the intermediate @xmath69 meson ; @xmath92 , @xmath93 , and @xmath94 stand for the 4-momentum , mass , and energy of the nucleon ; @xmath95 and @xmath96 the spin projection of the nucleon and the @xmath3 resonance ; @xmath97 and @xmath98 stand for the 4-momentum and energy of the final two pions , respectively . in our calculation , we use @xmath99 = 2.91 as the same as that used in ref . @xcite .
there is no information for the coupling constant of the @xmath60 vertex .
we determine it from the @xmath100 reaction .
we assume that the near threshold @xmath0 production in @xmath1 collisions is through the intermediate excitation of the sub-@xmath28-threshold @xmath3 resonance .
then , by comparing the theoretical total cross sections of @xmath101 reaction with experimental data , we can extract the @xmath60 coupling constant . in fig [ pipdiagram ]
, we show the s - channel diagram for the @xmath10 reaction , the intermediate excitation is a sub-@xmath102-threshold @xmath3 resonance . following the feynman rules and with the above formula
, we can obtain the invariant amplitude @xmath103 of the @xmath10 reaction , @xmath104 with @xmath105 , @xmath106 , @xmath107 , the spin projection of the @xmath0 meson and the nucleon , respectively .
the form factor for @xmath3 resonance , @xmath108 , is taken similar as in refs .
@xcite @xmath109 with @xmath110 = 2.0 gev .
@xmath111 is the propagator of the @xmath3 resonance , which can be written in a breit - wigner form @xcite , @xmath112 here @xmath113 is the energy dependent total width of the @xmath3 resonance .
according to pdg @xcite , the dominant decay channels for the @xmath3 resonance are @xmath114 and @xmath24 , so we take @xmath115 where @xmath116 is the following two - body phase space factor , @xmath117 from the amplitude , we can easily obtain the total cross sections of the @xmath10 reaction as functions of the excess energies . by adjusting the @xmath60 coupling constant , we can compare the theoretical results with the experimental data .
theoretical results with @xmath118 = 0.13 are compared with the experimental data in fig .
[ piptcs ] , we find an excellent agreement between our results and the experimental data .
contributions from the u - channel @xmath119 exchange and @xmath69-meson exchange between the pion and the proton are also checked and are found to be negligible . for @xmath100 reactions .
the experimental data are from ref .
, width=432,height=336 ] with experimental mass ( 1535 mev ) , width ( 150 mev ) , branching ratios of the @xmath3 @xcite and the total cross sections of the @xmath10 reaction , we obtain all the coupling constants as listed in table .
[ tab1 ] .
.[table ] relevant @xmath3 parameters . [ cols="^,^,^,^",options="header " , ] [ tab1 ] for the @xmath7 reaction , the full invariant amplitude in our calculation is composed of three parts corresponding to the @xmath3 resonance production from @xmath4 , @xmath5 and @xmath6-meson exchanges , respectively .
@xmath120 each amplitude can be obtained straightforwardly with the effective couplings and following the feynman rules . here
we give explicitly the amplitude @xmath121 , as an example , @xmath122 where @xmath107 is the spin projection of the @xmath0 meson ; @xmath123 and @xmath124 represent the spin projection and 4-momentum of the two initial and two final protons , respectively .
@xmath125 is the pion meson propagator , @xmath126 the final - state - interaction(fsi ) enhancement factor in the @xmath127 di - proton state are taken into account by means of the general framework based on the jost function formalism @xcite with latexmath:[\[\begin{aligned } where @xmath34 is the internal momentum of @xmath14 subsystem , and the @xmath129 and @xmath130 are related to the scattering parameters via @xmath131 with @xmath129 = -20.5 mev/@xmath132 and @xmath130 = 166.7 mev/@xmath132 @xcite ( i.e. @xmath133 = -7.82 fm and @xmath134 = 2.79 fm ) in the present study . then the calculations of the differential and total cross sections are straightforward , @xmath135 with the flux factor @xmath136 since the relative phases among different meson exchanges in the amplitude of eq .
( [ amp ] ) are not known , the interference terms are ignored in our concrete calculations .
with the formalism and ingredients given above , the total cross section versus excess energy @xmath137 for the @xmath138 reaction is calculated by using a monte carlo multi - particle phase space integration program .
it is known that the near - threshold production of the @xmath5 meson in @xmath139 reaction is thought to occur predominantly via the excitation of the @xmath3 resonance .
however , the excitation mechanism of the @xmath3 resonance in proton - proton collisions is currently still debated .
for example , batini et al .
@xcite and nakayama @xcite have found that the @xmath140 and @xmath5-meson exchanges between two protons play dominant roles for the excitation of the @xmath3 resonance
. however , gedalin et al .
@xcite and fldt and wilkin @xcite have found that the @xmath69-meson exchange is the dominant excitation mechanism of the @xmath3 resonance . here
the @xmath4 , @xmath5 and @xmath6-meson exchanges for @xmath3 excitation are all considered . by using the formalism and ingredients described in past section we first study the roles of different meson exchanges in the @xmath7 reaction .
our calculated results are shown in fig .
[ tcs ] together with the experimental data .
the double - dotted - dashed , dotted and dashed - dotted curves stand for contributions without the @xmath14 final - state - interaction ( fsi ) from @xmath4 , @xmath5 and @xmath6-meson exchanges , respectively . a simple summation of them are shown by the dashed line .
one can see that the contribution from the t - channel @xmath4 meson exchange is dominant to the @xmath7 reaction in our model .
the @xmath6-meson exchange has a significant contribution to this reaction , while the contribution from the @xmath5-meson exchange is negligible . from fig .
[ tcs ] we can see that our theoretical result without the @xmath14 fsi agrees well with the experimental data at excess energy @xmath137 = 83.0 mev .
however , at lower excess energies such as : @xmath137 = 18.5 , 34.5 mev , the calculated total cross sections are lower than the data by a factor of more than 4 .
it is known that the proton - proton fsi plays an important role for the near threshold meson production in proton - proton collisions .
we also include the effect of the @xmath127 @xmath14 fsi by using the jost - function method @xcite in our calculation , the results are shown in fig .
[ tcs ] by the solid line which can reproduce the anke total cross section data well . ) for the @xmath141 reaction from present calculation are compared with experimental data @xcite .
the double dotted - dashed , dotted , dashed - dotted and dashed curves stand for contributions from @xmath4 , @xmath5 , @xmath6-meson exchanges and their simple sum , respectively .
solid line corresponds to the results with the @xmath127 @xmath14 fsi.,width=403,height=288 ] the momentum , angular distributions of the @xmath0 meson and the @xmath142 invariant mass spectrum for the @xmath7 reaction at excess energy @xmath137 = 18.5 mev and 83.0 mev are also calculated . in fig .
[ phi185 ] we present our calculated results at excess energy @xmath137 = 18.5 mev together with experimental data from the anke group .
differential cross sections as a function of the c.m .
momentum of the outgoing proton are presented in the upper left panel .
the upper right panel is the angular distribution of the @xmath0 meson in the total proton - proton c.m .
the dashed lines are pure phase space distributions , while , the solid lines are full calculations from our model with the @xmath127 @xmath14 fsi enhancement factor . by comparing with the data , we find that the @xmath14 fsi plays an important role .
our model can explain the experimental data well .
in the lower part of fig .
[ phi185 ] the momentum distribution of the @xmath0 meson and the invariant mass spectrum of the outgoing proton and the @xmath0 meson are shown .
reaction at the excess energy @xmath137 = 18.5 mev compared with the anke data @xcite and phase space distribution ( dashed lines ) .
the upper left panel is the momentum distribution of the outgoing proton .
the upper right panel is the angular distribution of the @xmath0 meson in the total c.m .
frame ; the lower left panel is the distribution of the c.m .
momentum of the @xmath0 meson ; the lower right panel is the invariant mass spectrum of the outgoing proton and the @xmath0 meson .
, width=432,height=364 ] reaction at the excess energy @xmath137 = 83 mev compared with the disto data @xcite .
the dashed line reflects pure phase space , while the solid lines , which includes the amplitudes but without the @xmath14 fsi.,width=432,height=364 ] in fig .
[ phi83 ] , we present our calculated differential distributions at excess energy @xmath137 = 83.0 mev together with experimental data from the disto group . from our calculation
we find that there is no need to consider the @xmath14 fsi at this energy .
an excellent agreement between our model calculation and the experimental data both in shapes and magnitudes can be achieved without taking the @xmath14 fsi into account .
this is consistent with anke findings at @xmath137 = 75.9 mev .
the phenomena may suggest that at excess energy about 80 mev the contribution from @xmath14 higher partial waves has already overtaken the @xmath127 partial wave as the dominant contribution and the fsi becomes not important . in our calculation
we only include the contribution of the @xmath3 in the intermediate state . in previous calculations @xcite , the @xmath143 through t - channel @xmath69 exchange and/or sub - threshold nucleon pole contributions
are assumed to be dominant .
however these contributions are very sensitive to the choice of off - shell form factors for the t - channel @xmath69 exchange and the @xmath144 couplings and can be reduced by orders of magnitude within the uncertainties of these ingredients .
considering the ample evidence for large coupling of the @xmath3 to the strangeness @xcite and the @xmath3 resonance is closer than the nucleon pole to the @xmath28 threshold , it is more likely that the @xmath3 plays dominant role for near threshold @xmath0 production from @xmath145 and @xmath14 collisions instead of the nucleon pole or the ozi suppressed @xmath146 coupling .
our calculation with the @xmath3 domination reproduces energy dependence of the @xmath147 and @xmath148 cross sections better than previous calculations .
the significant coupling of the @xmath3 resonance to @xmath11 may be the real origin of the significant enhancement of the @xmath0 production from @xmath145 and @xmath14 reactions over the naive ozi - rule predictions .
this makes it difficult to extract the properties of the strangeness in the nucleon from these reactions proposed by j.ellis et al @xcite .
there are also some suggestions @xcite for possible existence of an @xmath2 bound state just below the @xmath2 threshold .
however , contribution of such bound state with width less than 100 mev will give a much sharper dropping structure for the @xmath147 cross section at energies near threshold .
if such @xmath2 bound state does exist , it should have weak coupling to @xmath114 and only gives small contribution to the @xmath147 reaction .
in this paper , the near threshold @xmath0 meson productions in proton - proton and @xmath1 collisions are studied with an effective lagrangian approach .
we assume that the production mechanism is due to the excitation of the sub-@xmath2-threshold @xmath3 resonance following @xmath4 , @xmath5 and @xmath6-meson exchanges between two protons .
@xmath149 , @xmath150 and @xmath151 coupling constants ( except @xmath152 ) and form factors are taken from the bonn potential model .
@xmath153 , @xmath58 and @xmath154 coupling constants are determined from the partial decay widths of the @xmath3 resonance .
the @xmath60 coupling constant is deduced from a fit to the experimental total cross sections of the @xmath101 reaction near threshold with the @xmath3 resonance model .
we find that the @xmath3 resonance has a significant coupling to @xmath11 ( @xmath155 = 0.13 ) .
the total reaction cross sections and differential distributions of the near threshold @xmath7 reaction are calculated with the @xmath3 resonance model without adjustable parameter .
our theoretical calculation agrees quite well with experiments near threshold .
we find that the contribution from the t - channel @xmath4 meson exchange is dominant to the @xmath7 reaction .
the significant coupling of the @xmath3 resonance to the @xmath28 channel together with the earlier findings of large couplings of the @xmath3 resonance to the @xmath24 , @xmath27 and @xmath18 channels @xcite gives a coherent picture that there is a large component of strangeness in the @xmath3 resonance as expected by various theoretical approaches @xcite .
it also gives a natural explanation for the significant enhancement of the @xmath0 production from @xmath145 and @xmath14 reactions over the naive ozi - rule predictions .
for a better understanding of the dynamics of these reactions , more experimental data at other excess energies with dalitz plots and angular distributions are desired .
* acknowledgement * + we would like to thank feng - kun guo , bo - chao liu , colin wilkin , feng - quan wu and hai - qing zhou for useful discussions .
this work is partly supported by the national natural science foundation of china under grants nos .
10435080 , 10521003 and by the chinese academy of sciences under project no .
kjcx3-syw - n2 .
99 c. hanhart , phys .
rept . * 397 * , 155 ( 2004 ) .
zou , int . j. mod
a * 21 * , 835 ( 2006 ) and references therein .
s. okubo , phys . lett . * 5 * , 165 ( 1963 ) ; + g. zweig , cern report th-412 ( 1964 ) ; + j. iizuka , prog .
phys . suppl . *
38 * , 21 ( 1966 ) .
f. balestra et al.(disto collaboration ) , phys .
lett . * 81 * , 4572 ( 1998 ) ; phys .
c * 63 * , 024004 ( 2001 ) .
m. hartmann et al . , nucl .
a * 755 * , 459c ( 2005 ) .
a. sibirtsev , j. haidenbauer and u .- g .
meisser , eur .
j. a * 27 * , 263 ( 2006 ) ; + a. sibirtsev and w. cassing , eur .
j. a * 7 * , 407 ( 2000 ) .
titov , b. kmpfer and b.l .
reznik , eur .
phys . j. a * 7 * , 543 ( 2000 ) .
k. nakayama , nucl - th/0108032 ; + k. tsushima and k. nakayama , phys .
c * 68 * , 034612 ( 2003 ) .
l.p . kaptari and b. kmpfer , eur .
j. a * 23 * , 291 ( 2005 ) .
m. hartmann et al . , phys . rev .
lett . * 96 * , 242301 ( 2006 ) .
liu and b.s .
zou , phys .
* 96 * , 042002 ( 2006 ) ; + b.c .
liu and b.s .
zou , phys .
lett . * 98 * , 039102 ( 2007 ) ; + b.c .
liu and b.s .
zou , commun .
46 * , 501 ( 2006 ) .
g. penner and u. mosel , phys .
c * 66 * , 055211 ( 2002 ) ; ibid .
c * 66 * , 055212 ( 2002 ) ; + v. shklyar , h. lenske and u. mosel , phys .
c * 72 * , 015210 ( 2005 ) . b. julia - diaz , b. saghai , t.s.h .
lee and f. tabakin , phys .
c * 73 * , 055204 ( 2006 ) .
tran et al . , phys .
b445 * , 20 ( 1998 ) ; + k.h .
glander et al . , euro .
j. * a19 * , 251 ( 2004 ) .
mcnabb et al . , phys .
c * 62 * , 042201(r ) ( 2004 ) .
t. inoue , e. oset and m.j .
vicente vacas , phys .
c * 65 * , 035204 ( 2001 ) ; + n. kaiser , t. waas and w. weise , nucl .
a * 612 * , 297 ( 1997 ) .
m. dugger et al .
lett . * 96 * , 062001 ( 2006 ) ; erratum - ibid .
* 96 * , 169905 ( 2006 ) .
r. machleidt , k. holinde and c. elster , phys . rep . * 149 * , 1 ( 1987 ) ; + r. machleidt , adv .
* 19 * , 189 ( 1989 ) ; + r. brockmann and r. machleidt , phys .
c * 42 * , 1965 ( 1990 ) .
k. tsushima , s.w . huang and a. faessler ,
b * 337 * , 245 ( 1994 ) ; + k. tsushima , a. sibirtsev and a.w .
thomas , phys .
b * 39 * , 29 ( 1997 ) ; + k. tsushima , a. sibirtsev , a.w .
thomas and g.q .
li , phys .
c * 59 * , 369 ( 1999 ) , erratum - ibid .
c * 61 * , 029903 ( 2000 ) .
a. sibirtsev and w. cassing , nucl - th/9802019 ; + a. sibirtsev , k. tsushima , w. cassing and a. w. thomas , nucl .
a * 646 * , 427 ( 1999 ) .
w. grein and p. kroll , nucl .
a * 338 * 332 , ( 1980 ) ; + m. kirchbach and l. tiator , nucl .
a * 604 * , 385 ( 1996 ) ; + s.l .
zhu , phys .
c * 61 * , 065205 ( 2000 ) .
g. fldt and c. wilkin , physica scripta , * 64 * , 427 ( 2001 ) .
b.s . zou and f. hussain , phys .
c * 67 * , 015204 ( 2003 ) .
particle data group , j. phys .
g * 33 * , 1 ( 2006 ) .
vrana , s.a .
dytman and t.s.h .
lee , phys . rept .
* 328 * , 181 ( 2000 ) .
bugg and b.s .
d * 55 * , 1421 ( 1997 ) .
t. feuster and u. mosel , phys .
c * 58 * , 457 ( 1998 ) ; + t. feuster and u. mosel , phys . rev .
c * 59 * , 460 ( 1999 ) .
w.h . liang , p.n .
shen , j.x .
wang and b.s .
zou , j. phys .
g * 28 * , 333 ( 2002 ) .
a. baldini , v. flamino , w.g .
moorhead and d.r.o .
morrison , @xmath156-@xmath157@xmath158 , @xmath159 @xmath160 @xmath161 @xmath162 @xmath163 @xmath164 @xmath165 @xmath166 @xmath167 , vol.*12 * , ed . by h. schopper , springer - verlag(1988 ) ,
@xmath168 @xmath169 @xmath170 @xmath171 @xmath172 @xmath173 @xmath174 .
j. gillespie , @xmath175 @xmath176 , holden - day advanced physics monographs , ( 1964 ) .
m. batini , a. svarc and t.s.h .
lee , physica scripta , * 56 * , 321 ( 1997 ) .
k. nakayama , j. speth and t.s.h .
lee , phys .
c * 65 * , 045210 ( 2002 ) ; + k. nakayama , j. w. durso , j. haidenbauer , c. hanhart , and j. speth , phys .
c 60 , 055209 ( 1999 ) .
g. gedalin , a. moalem and l. razdolskaja , nucl .
a * 634 * , 368 ( 1998 ) .
l.hannelius and d.o.riska , phys .
c * 62 * , 045204 ( 2000 ) .
j. ellis , m. karliner , d.e .
kharzeev and m.g .
sapozhnikov , nucl .
phys . a * 673 * , 256 ( 2000 ) .
h. gao , t.s.h .
lee and v. marinov , phys .
c * 63 * , 022201 ( 2001 ) .
f. huang , z.y .
zhang and y.w .
yu , phys .
c * 73 * , 025207 ( 2006 ) .
a. zhang et al . , high energy phys .
* 29 * , 250 ( 2005 ) . | the near threshold @xmath0 meson production in proton - proton and @xmath1 collisions is studied with the assumption that the production mechanism is due to the sub-@xmath2-threshold @xmath3 resonance .
the @xmath4 , @xmath5 and @xmath6-meson exchanges for proton - proton collisions are considered .
it is shown that the contribution to the @xmath7 reaction from the t - channel @xmath4 meson exchange is dominant . with a significant @xmath8 coupling ( @xmath9 = 0.13 ) ,
both @xmath7 and @xmath10 data are very well reproduced .
the significant coupling of the @xmath3 resonance to @xmath11 is compatible with previous indications of a large @xmath12 component in the quark wave function of the @xmath3 resonance and may be the real origin of the significant enhancement of the @xmath0 production over the naive ozi - rule predictions . |
galactic cosmic rays ( gcr ) originate outside the solar system .
they comprise protons ( 85@xmath0 ) , alpha particles ( 14@xmath0 ) and heavy nuclei .
interaction with large - scale ordered magnetic field causes the gradient- and curvature - drift motion of gcr in the inner heliosphere , while the interaction with the irregular ( stochastic ) field component results in the pitch angle scattering of gcr .
the scale sizes for the two effects , drifts which depend on variations in the mean field on the order of the heliocentric radial distance , and diffusion which depends on irregularities of sizes comparable to the particles gyro - radii , are quite distinct ( see , for instance @xcite,@xcite and references therein ) . until recently space instruments have lacked the combination of large geometrical factor and good mass resolution required to address the question of the variation on _ short _ time scales of individual nuclides other than @xmath1 and @xmath2 . in present letter
we report on an investigation of the flux fluctuations of gcr nuclei ( carbon - @xmath3 : @xmath4 ) based on the data obtained by instrumentation carried aboard the advanced composition explorer ( ace ) spacecraft for 2000 year .
the intensities of these low energy particles have been continuously monitored by the cosmic ray isotope spectrometr ( cris ) . in order to get away from the effects of the earth s magnetic field , the ace spacecraft orbits at the l1 libration point which is a point of earth - sun gravitational equilibrium about 1.5 million km from earth and 148.5 million km from the sun .
the large collecting power and high resolution of the cris instrument allow us to investigate modulation of gcr flux on short time scales ( beginning from 1 hour ) and in different energy ranges .
current theories of parallel and perpendicular diffusion are fairly well accepted for high energy particles .
however , there are important outstanding issues pertaining to diffusion and transport of charged particles at medium to low energies .
the particles at low energies are believed interact directly with the steepened dissipation range that sets in near the ion inertial scale .
dynamical scattering theory suggests that these particles may be sensitive to ( 1 ) the dynamical decorrelation mechanism that operates at these scales ; ( 2 ) the `` geometry '' of the fluctuations in the dissipation range ; and ( 3 ) the spectral distribution of power at the small scales @xcite .
+ multiscaling properties of cosmic rays have been already discussed in the light of the data obtained by neutron monitors with cut - off rigity 1.09 gev ( see , for instance , @xcite,@xcite and references therein ) as well as modulation of the cosmic rays by solar wind turbulence and magnetic field intermittency @xcite-@xcite . however , as far as we know , multiscaling analysis of the gcr heavy nuclei data obtained by satellite instruments is performed here for the first time .
this analysis reveals a strong ( turbulent - like ) intermittency of the gcr dissipation rate and dependence of the intermittency type from the energy ranges ( not to be confused with the hep multiplicity intermittency @xcite,@xcite ) .
differential flux of the gcr : @xmath5 , as measured by cris , is proportional to the normal component of the velocity field of the gcr : @xmath6 , and to their local concentarion : @xmath7 , @xmath8 fluctuations of @xmath5 with time are then determined by corresponding fluctuations of @xmath6 and @xmath7 .
dissipation of passive admixture concentration in fluid turbulence is characterized by a `` gradient '' measure @xcite-@xcite : @xmath9 where @xmath10 is a subvolume with space - scale @xmath11 ( for detail justification of this measure see handbook @xcite : p. 381 and
further ) . scaling law of this measure moments , @xmath12 is an important characteristic of the dissipation rate field just in _ inertial _ interval of turbulence ( see , for instance , @xcite-@xcite ) .
analogous measure is used to characterize also dissipation rate of turbulent velocity ( or kinetic energy ) in the inertial interval of scales @xcite-@xcite . for turbulent flows the taylor hypothesis
is generally used to interpret the data @xcite,@xcite,@xcite .
this hypothesis states that the intrinsic time dependence of the wavefields ( @xmath6 and @xmath7 ) can be ignored when the turbulence is convected past the probes at nearly constant speed . with this hypothesis
, the temporal dynamics should reflect the spatial one , i.e. the fluctuating velocity ( concentration ) field measured by a given probe as a function of time , @xmath13 is the same as the velocity @xmath14 where @xmath15 is the mean velocity and @xmath16 is the distance to a position `` upstream '' where the velocity is measured at @xmath17 . with the taylor hypothesis @xmath18
is replaced by @xmath19 and one can define gcr flux dissipation rate as : @xmath20 where @xmath21 and corresponding scaling of the dissipation rate moments as @xcite : @xmath22 substituting ( 1 ) into ( 4 ) one can estimate the gcr flux dissipation rate through characteristics of @xmath6 and @xmath7 .
we can consider two asymptotic regimes .
for sufficiently small gcr energies the flux dissipation rate has been dominated by the concentration dissipation : @xmath23^p \rangle \sim \tau^{-\mu_p } \eqno{(6)}\ ] ] while for sufficiently large gcr energies the flux dissipation rate has been dominated by the energy ( velocity ) dissipation : @xmath24^p \rangle \sim \tau^{-\mu_p } \eqno{(7)}\ ] ]
as it was mentioned in introduction we will use the data collected by ace / cris instrument during 2000 year ( the year of maximum solar activity ) .
we will consider carbon nuclei @xmath3 ( @xmath4 ) , due to carbon is the lightest abundant nuclei ( after @xmath1 and @xmath2 ) in the ace / cris collection . as we shall see
this fact allow us consider the both asymptotes mentioned in previous section .
figure 1a shows scaling of the gcr flux dissipation rate moments @xmath25 ( 4)-(5 ) for the c - nuclei with energies from energy range : 56.3 - 73.4 mev / nucleon ( the lowest energy range in the ace / cris collection for carbon ) .
figure 1b shows the scaling exponents @xmath26 ( circles ) extracted from figure 1a ( as slopes of the straight lines , cf ( 5 ) ) .
the solid curve in figure 1b corresponds to the intermittency exponents @xmath26 obtained for the _ inertial_-convective region of a passive admixture concentration in a laboratory turbulent air flow @xcite . to support the striking correspondence between the two data sets multiscaling
let us calculate also the extended self - similarity ( ess @xcite ) exponents , @xmath27 , extracted from equation : @xmath28 the ess of type ( 8) usually has clearer scaling form than ordinary scaling ( 5 ) and covers a wider range of scales ( see @xcite for a review of ess and for examples ) .
figure 2a shows the ess of the gcr flux dissipation rate moments @xmath25 ( 8) using log - log scales .
the straight lines ( the best fit ) are drawn to indicate the ess ( 8) .
figure 2b shows the ess exponents @xmath27 ( circles ) extracted from figure 2a ( as slopes of the straight lines , cf ( 8) ) .
the solid curve in figure 2b corresponds to the intermittency exponents @xmath27 obtained for the _ inertial_-convective region of passive admixture concentration in different laboratory turbulent flows @xcite .
now let us turn to the data corresponding to energy range with the highest for the carbon nuclei energies observed by ace / cris .
figure 3a shows scaling of the gcr flux dissipation rate moments @xmath25 ( 4)-(5 ) for the c - nuclei with energies from energy range : 183.1 - 198.7 mev / nucleon .
figure 3b shows the scaling exponents @xmath26 ( circles ) extracted from figure 3a ( as slopes of the straight lines , cf ( 5 ) ) .
the solid curve in figure 3b corresponds to the intermittency exponents @xmath26 calculated using the she - leveque model @xcite , which is in very good agreement with the data for _ velocity _ ( kinetic energy dissipation ) field intermittency obtained in _ inertial _
interval for isotropic fluid turbulence .
figures 4a and 4b shows corresponding extended self - similarity ( ess ) properties @xmath29 observed for the data . again , as in figure 3b , correspondence to the fluid turbulence ( energy dissipation rate - the solid curve in figure 4b ) is very good .
the interval of time scales under consideration is 1 - 30 hours .
it seems to be useful to compare the observed properties of gcr flux dissipation rate with relevant properties of the local interplanetary magnetic field .
figure 5 shows energy spectrum of 3d magnetic field measured by ace / mag magnetometer in the same time intervals as the ace / cris data .
the slope -5/3 indicates kolmogorov - like scaling ( cf @xcite ) .
let us recall that the turbulent fluid dissipation rates used for comparison in figures 1 - 4 were obtained just for inertial ( inertial - convection ) interval of scales where the kolmogorov - like scaling should to be expected @xcite,@xcite .
the results of previous section seems to be in agreement with the qualitative estimates made in section 2 : for relatively small energies the flux dissipation rate behaves as one dominated by the gcr particles concentration dissipation rate , whereas for relatively large energies the gcr flux dissipation rate seems to be dominated by velocity ( kinetic energy ) dissipation . since cris instrument measures the differential flux ( @xmath5 ) only , but not the velocity of gcr ( @xmath6 ) and their concentration ( @xmath7 ) separately ( cf ( 1 ) ) we have no idea about magnitudes of the fluctuations of @xmath6 and @xmath7 themselves .
therefore existence of the two asymptotes ( 6 ) and ( 7 ) is still a pure phenomenology ( for the intermediate energies the scaling is deformed by competition between the two different mechanisms of the flux dissipation ) .
the same reason ( unknown @xmath6 and @xmath7 for the gcr ) makes it impossible to perform any theoretical estimates for the scales of the gcr flux dissipation rate multiscaling . although correspondence between the motions in mhd and hydro cases has been reported earlier ( see , for instance , very recent paper @xcite and references therein ) it is much more difficult to understand such detail correspondence between cosmic rays and fluid turbulence . indeed , as it follows from previous two sections the observed gcr particles dissipation rate exhibits intermittent properties of a passive admixture convected by turbulent motion of a classic ( isotropic , incompressible ) non - magnetic fluid .
while the solar wind ( including the interplanetary magnetic field - imf ) is certainly very different from such a fluid . moreover , interaction of the electrically charged gcr particles with the solar wind and the magnetic field is expected to have strong resonance features ( see , for instance , @xcite-@xcite,@xcite and references therein ) .
therefore , even the fact that many characteristics of solar wind and imf turbulence are known to be similar to those of turbulence of classic non - magnetic fluids ( see @xcite-@xcite and references therein ) the presumably
_ resonance_-like interactions of gcr particles with solar wind and imf seems to be non - consistent with the observed here picture , that request reconsideration of our approach to stochastic _
short_-term convection of the low - energy gcr particles ( heavy nuclei ) in the heliosphere and their _ short_-term interactions with solar wind and imf .
this reconsideration represents a serious challenge for a wide scientific community : both for astrophysicists and for plasma experts .
+ the author is grateful to ace / cris and to ace / mag instrument teams as well as to the ace science center for providing the data .
numerous discussions on the subject with k.r .
sreenivasan as well as the referees comments and suggestions were very useful for this investigation .
fisk and n.a .
schwadron , j. geophys .
* 100 * , 7865 ( 1995 ) .
alania , acta physica polonica b , * 33 * , 1149 ( 2002 ) .
matthaeus , in `` sun - earth connection guest investigator program : abstracts of awarded proposals '' ( nra-01-oss-01 ) .
k. kudela and d. venkatesan , nucl .
b ( suppl .
a ) * 39 * , 127 ( 1995 ) .
a. bershadskii , j. phys .
g , * 27 * , l61 ( 2001 ) .
horbury , in " plasma turbulence and energetic particles , ed . by m. ostrowski and r.shlickeiser , ( cracow , 1999 ) , p. 28 . b. chandran , phys .
lett . , * 85 * , 4656 ( 2001 ) .
r. schlickeiser , cosmic ray astrophysics , ( springer - verlag , berlin 2002 ) .
j. cho , a. lazarian and e.t .
vishniac , astro - ph/0205286 ( 2002 ) .
o. stawicki , s.p .
gary , h. li , j. geophys res . - space physics , a5 , * 106 * , 8273 ( 2001 ) .
goldstein , astrophys .
space sci . * 227 * , 349 ( 2001 ) e. marsch , and s. liu , ann . ,
geophysicae , * 11 * , 227 ( 1993 ) .
horbury aned a. balogh , nonlin .
geophys . , * 4 * 185 ( 1997 ) .
a. bialas and r. peschanski , nucl .
b , * 273 * , 703 ( 1986 ) ; * 308 * , 857 ( 1988 ) .
a. bershadskii , phys .
c , * 59 * , 364 ( 1999 ) .
a.c monin and a.m.yaglom , statistical fluid mechanics , vol .
2 ( mit press , cambridge , 1975 ) .
sreenivasan , annu .
fluid mech .
* 23 * , 539 ( 1991 ) .
k.r . sreenivasan and r.a .
antonia , annu .
fluid mech , * 29 * , 435 ( 1997 ) .
r. benzi , l. biferale , s. ciliberto , m.v .
struglia , r. tripiccione , physica d , * 96 * , 162 ( 1996 ) .
she and e. leveque , phys .
lett . , * 72 * , 336 ( 1994 )
. j. cho , a. lazarian and e.t .
vishniac , apj , * 564 * , 291 ( 2002 ) .
figure 1a .
the gcr flux dissipation rate moments @xmath25 against @xmath30 for the c - nuclei with energies from energy range : 56.3 - 73.4 mev / nucleon .
the time interval @xmath30 is measured in hours , whereas the gcr differential flux @xmath5 is measured in @xmath31 .
log - log scales are chosen in the figure for comparison with scaling equation ( 5 ) .
the straight lines ( the best fit ) are drawn to indicate the scaling .
the upper data sets correspond to larger @xmath32 + figure 1b .
the scaling exponents @xmath26 ( circles ) extracted from figure 1a .
the solid curve corresponds to the intermittency exponents @xmath26 obtained for the inertial - convective region of a passive admixture concentration dissipation rate in a laboratory turbulent air flow @xcite .
+ figure 2a .
the ess of the gcr flux dissipation rate moments @xmath25 against @xmath33 in log - log scales ( 8) for energy range : 56.3 - 73.4 mev / nucleon . the straight lines ( the best fit ) are drawn to indicate the ess ( 8) . the upper data sets correspond to larger @xmath32 .
+ figure 2b .
the ess exponents @xmath27 ( circles ) extracted from figure 2a .
the solid curve corresponds to the intermittency exponents @xmath27 obtained for the inertial - convective region of passive admixture concentration dissipation rate in different fluid turbulent flows @xcite .
+ figure 3a .
the gcr flux dissipation rate moments @xmath25 ( 4)-(5 ) against @xmath30 for the c - nuclei with energies from energy range : 183.1 - 198.7 mev / nucleon . + figure 3b .
the scaling exponents @xmath26 ( circles ) extracted from figure 3a .
the solid curve in figure 3b corresponds to the intermittency exponents @xmath26 calculated using the she - leveque model for isotropic fluid turbulence @xcite .
+ figure 4a .
the ess of the gcr flux dissipation rate moments @xmath25 against @xmath33 in log - log scales ( 9 ) for energy range : 183.1 - 198.7 mev / nucleon . the straight lines ( the best fit )
are drawn to indicate the ess ( 9 ) .
+ figure 4b .
the ess exponents @xmath27 ( circles ) extracted from figure 4a .
the solid curve corresponds to the intermittency exponents @xmath27 calculated using the she - leveque model .
+ figure 5 .
energy spectrum of 3d magnetic field measured by ace / mag magnetometer in the discussed range of the time - scales . | multiscaling analysis of differential flux dissipation rate of galactic cosmic rays ( carbon nuclei ) is performed in the energy ranges : 56.3 - 73.4 mev / nucleon and 183.1 - 198.7 mev / nucleon , using the data collected by ace / cris spacecraft instrument for 2000 year .
the analysis reveals strong ( turbulence - like ) intermittency of the flux dissipation rate for the short - term intervals : 1 - 30 hours .
it is also found that type of the intermittency can be different in different energy ranges . |
recently there was an experimental development in fabrication of atomically thin graphene , or single - layer graphite , that enables us to access its exotic electronic properties.@xcite the magnetotransport was measured and the integer quantum hall effect was observed.@xcite in the experiments , a multilayer that contains a few graphene sheets is also available.@xcite the electronic structure of a bilayer graphene was studied theoretically and the spectrum was found to be essentially different from that of a monolayer.@xcite the purpose of this paper is to study transport properties of the bilayer graphene .
a graphite monolayer has a @xmath0-linear , massless dirac - like spectrum and has long attracted theoretical interests as a `` relativistic '' problem in condensed matter physics , where @xmath0 is the absolute value of the wave vector .
theoretical studies of transport in such an exotic electronic structure have been given by several authors , where the conductivity with or without magnetic field,@xcite the hall effect,@xcite quantum corrections to the conductivity,@xcite and the dynamical transport@xcite are investigated .
the results show that the conductivity exhibits various singular behaviors in the vicinity of zero energy .
@xcite in bilayer graphene the energy dispersion includes both @xmath0-linear and @xmath0-square terms . in this paper , we calculate the diagonal conductivity in the absence of a magnetic field within a self - consistent born approximation ( scba ) .
a similar scba analysis has been applied for monolayer graphene.@xcite we shall find that the coexistence of @xmath0-linear and @xmath0-square dispersions leads to a crossover between transport properties similar to those of a monolayer graphene and to those of an ordinary two - dimensional metal as the fermi energy is changed .
the conductivity becomes nearly universal , @xmath1 per spin , in the case of large disorder .
the analysis is made for two different kinds of scatterers , short range and long range , where the former represents on - site random energies distributed on the carbon atoms and the latter a slowly varying random potential of the range much longer than the lattice constant but shorter than the typical electron wavelength . in sec .
ii , the effective hamiltonian and the resulting energy spectrum in a bilayer graphene are discussed , and model scatterers are introduced . the self - consistent born approximation is briefly described in sec .
explicit results are discussed for short - range scatterers in sec .
iv and for long - range scatterers in sec . v. a discussion and a brief summary are given in sec .
vi . [ cols="^,^ " , ] figure [ fig_disp ] shows the energy dispersion ( [ eq_e ] ) for @xmath2 . in the high - energy region @xmath3
, we have a single trigonally warped fermi line , which becomes closer to a circle on going to the higher energy as the @xmath0-square term dominates in the hamiltonian .
the equienergy line becomes convex for @xmath4 .
in the low - energy region @xmath5 , the fermi line splits into four separate pockets for each of @xmath6 and @xmath7 , one center part and three satellite parts that are located trigonally . in the vicinity of zero energy @xmath8 ,
the dispersion becomes linear in @xmath0 space with respect to the four fermi points , where the center pockets can be approximated as a circle with radius @xmath9 and the three satellites as ellipses with the longer and shorter radii @xmath10 and @xmath11 .
while splitting of the four fermi lines occurs only at very low energies , the trigonal warping extends to much higher energy as seen in fig .
[ fig_disp ] . the velocity operator @xmath12 has nonzero matrix elements between the states on the identical @xmath0 points @xmath13 written as @xmath14 with @xmath15 in the high - energy region @xmath3 , the mean square velocity averaged on the contour @xmath16 is given by @xmath17 where the first and second terms come from the @xmath0-square and @xmath0-linear terms in the dispersion , respectively . in the vicinity of zero energy @xmath8
, we have @xmath18 the density of states per spin is defined by @xmath19 where @xmath20 is the area of the system . in the high energy region
@xmath3 , this becomes @xmath21 with terms of the order of @xmath22 neglected . in the vicinity of zero energy @xmath8 , on the other hand ,
@xmath23 the density of states diverges logarithmically at @xmath24 due to the presence of saddle points in the dispersion . for the parameters mentioned above , @xmath25 with @xmath26 being the free - electron mass , @xmath27 , and @xmath28 mev .
the electron concentration corresponding to @xmath29 is @xmath30 @xmath31 , that to @xmath32 is @xmath33 @xmath31 , and that to @xmath34 is @xmath35 @xmath31 . as the typical electron concentration in the present system is @xmath36 @xmath31 ( @xmath37 ) or larger,@xcite the trigonal warping
is appreciable , but it is extremely hard to realize the situation where four fermi lines are well split from each other . the hamiltonian ( [ eq_heff ] ) is formally equivalent to that of the monolayer system , where the @xmath0-square term comes from a higher - order term in the @xmath38 approximation.@xcite however , the parameter @xmath39 for the monolayer becomes the order of @xmath40 and thus is larger than in the bilayer by the order of 1000 , and the @xmath0-square term gives only a small perturbation . in the bilayer , on the contrary , the @xmath0-square term becomes dominant in the dispersion at small energy @xmath32 .
note also that the hamiltonian ( [ eq_heff ] ) becomes invalid when the energy becomes as high as the antibonding states of the @xmath41-@xmath42 dimers .
the deviation from ( [ eq_heff ] ) seems to become appreciable around@xmath43 ( @xmath44 ev corresponding to@xmath45 @xmath31).@xcite the dispersion of the bilayer graphene is much closer to that of three - dimensional graphite , where we see a similar trigonal structure in the @xmath46 plane with a fixed wave number @xmath47 along the stacking direction.@xcite in terms of the eigenvector @xmath48 the amplitudes of the atomic orbitals at sites @xmath49 and @xmath50 are given by @xmath51 , \\
\psi_{b}(\vec{r}^{b } ) & = & { \omega_j \over \sqrt n } \phi_{b}^{j{\bf k}s } \exp[i(\vec{k}_j + \vec{k})\cdot\vec{r}^{b } ] , \end{aligned}\ ] ] for @xmath52 and @xmath7 , where @xmath53 , @xmath54 with @xmath55 , and @xmath56 is the number of unit cells in the system .
the dominant scatterers in the present system are not well known . in the following ,
we shall consider two kinds of model scatterers , short - range scatterers localized only on @xmath57 sites or on @xmath58 sites and long - range scatterers , the potential of which spreads over a certain length scale larger than the lattice constant @xmath59 . in each case
we assume that the disorder strength is weaker than the dimer coupling @xmath60 , so that we need to consider only the potential on @xmath58 and @xmath57 sites .
the effective hamiltonian for the disorder potential can be derived similarly as in monolayer graphene , if we identify the @xmath58 and @xmath57 sites in the bilayer with @xmath41 and @xmath57 in the monolayer.@xcite the matrix elements for the short - ranged potential are written as @xmath61 for scatterers at @xmath62 and @xmath63 , respectively , where @xmath64 and @xmath65 are the integrated intensities of the potential , i.e. , the coefficient of the @xmath66 potential . for long range disorder , we assume that the potential range is much larger than the lattice constant @xmath59 but smaller than the typical wavelength @xmath67 with @xmath0 being the wave number from the @xmath6 or @xmath7 point
. then we can neglect the matrix elements for intervalley scatterings between @xmath6 and @xmath7 , while those for intravalley can be written as @xmath68 where @xmath69 is the integrated intensity of the potential .
we have another possibility for the long - range disorder potential , where each of the scatterers is effective only in one layer .
this is modeled by a long - range potential that has amplitude only over either @xmath58 or @xmath57 sites .
the situation then becomes almost equivalent to the short - range case , the only difference being that @xmath6 and @xmath7 are decoupled , reducing the self - energy by a factor of 2 .
in the self - consistent born approximation , the self - energy of the disorder - averaged green s function @xmath70 is given by@xcite @xmath71 with @xmath72 , where @xmath73 represents the average over the impurity configurations .
in the present system the conductivity becomes isotropic in spite of the presence of strong trigonal warping .
it can be calculated by the kubo formula , @xmath74 , \ ] ] where @xmath75 and @xmath76 are the retarded and the advanced green s functions , respectively , with @xmath77 being the hamiltonian including the disorder potential .
this can be rewritten as @xmath78,\ ] ] with @xmath79 and @xmath80 satisfying @xmath81 in the scba , @xmath82 should be calculated in the ladder approximation . in the above and
hereafter we omit the summation over the spin degeneracy , so the actual conductivity should be multiplied by a factor of 2 . in the case of the present model scatterers , the self - energy and therefore the averaged green s function
become diagonal with respect to the wave number and the band index .
further , the self - energy is independent of the wave number and the band index , and thus is determined by the energy alone . as a result we have @xmath83 where @xmath84 is the self - energy .
for short - range scatterers , we assume that they are equally distributed to @xmath41 and @xmath57 sites with density @xmath85 and the identical mean square amplitude @xmath86 .
then , the self - energy is given by @xmath87 by substituting the summation over @xmath88 and @xmath89 with the integration in energy @xmath90 , we can rewrite this as @xmath91 the density of states is given by @xmath92 for the conductivity , we can show that the vertex correction vanishes in the short - range scatterers , or @xmath93 , and we obtain @xmath94 .
\label{eq_sigma}\end{aligned}\ ] ] in the second equality we have replaced the summation over @xmath88 and @xmath95 with the integral in @xmath96 .
we first look at the boltzmann limit by taking @xmath97 , which should be valid in the weak - disorder case satisfying @xmath98 . by assuming that @xmath99 is infinitesimal in ( [ eq_self ] )
, we can calculate @xmath99 explicitly as @xmath100 where @xmath101 is the lifetime and the dimensionless parameter @xmath102 is defined by @xmath103 for sufficiently high energy , @xmath104 showing that @xmath105 characterizes a typical disorder strength that smears out the fine low - energy structure due to the @xmath0-linear term .
this @xmath102 is the same as that defined for a monolayer graphene in ref . and as @xmath106 in ref .
. for the conductivity ,
we take only the terms with @xmath107 and @xmath108 as a dominant contribution in ( [ eq_sigma ] ) , and obtain a familiar form , @xmath109 here @xmath110 is the fermi - surface average of the diagonal matrix element of @xmath111 and use has been made of the approximation @xmath112 with the use of ( [ eq_vsq_high ] ) the conductivity for the higher energy @xmath113 becomes @xmath114 the term proportional to @xmath115 results from the @xmath0-square term of the dispersion and is rewritten as @xmath116 as in the usual two - dimensional metal .
the term independent of @xmath117 is a correction due to the trigonal warping caused by the @xmath0-linear term , which never vanishes even as @xmath118 . in the low - energy region @xmath119 , we have @xmath120 the energy - independent conductivity is essentially the same as that of monolayer graphene.@xcite s . for the boltzmann conductivity in the lower panel we used the expression ( [ eq_sigma_weak_high ] ) which is valid for @xmath121 . in the lower panel
the approximate result given by eq .
( [ eq_sigma_scba ] ) is also shown.,width=302 ] in the case of large disorder @xmath122 , the fine structure in the density of states disappears completely ( see below ) and therefore the imaginary part of the self - energy becomes independent of energy and the real part becomes negligible .
thus , we always have @xmath123 with @xmath124 giving @xmath125 we can calculate the conductivity ( [ eq_sigma ] ) by substituting the high - energy expansions ( [ eq_vsq_high ] ) and ( [ eq_rho_high ] ) , and obtain @xmath126 + \frac{e^2}{\pi^2\hbar}\frac{1}{w } , \label{eq_sigma_scba}\ ] ] with @xmath127 where terms of the order of @xmath128 are neglected .
the first term of ( [ eq_sigma_scba ] ) comes from the @xmath0-square term in the dispersion and the second term is a correction due to the @xmath0-linear term .
the conductivity goes to the boltzmann limit ( [ eq_sigma_weak_high ] ) for @xmath129 , while for @xmath130 it becomes @xmath131 it is interesting that the conductivity becomes universal , i.e. , @xmath132 , and never vanishes in the limit of the large disorder . at a rough estimate , we can derive this expression by putting the uncertainty relation @xmath133 in ( [ eq_sigma_weak_high ] ) . in sec .
vi , this universal conductivity will be reconsidered in terms of einstein s relation .
the self - consistent equation ( [ eq_self ] ) can be solved easily by a numerical iteration .
figure [ fig_dnst ] shows the density of states calculated for several disorder strengths .
we notice that the logarithmic divergence present in @xmath134 at @xmath135 is smeared out very easily , and the structure in the vicinity of @xmath136 due to the @xmath0-linear dispersion disappears around @xmath137 and is almost unrecognizable already for @xmath138 . the corresponding plot for the conductivity is shown in fig .
[ fig_cond_short ] along with the boltzmann limit . in weak disorder @xmath139
we observe a sharp dip at zero energy .
this is the analog of monolayer graphene,@xcite showing that @xmath0-linear dispersion around zero energy remains intact at small disorder .
further discussion of the asymptotic value at @xmath140 will be given in sec.vi .
the boltzmann conductivity drops to zero at @xmath141 , as the velocity vanishes at the saddle points in the dispersion , but this singularity disappears due to the finite density of states . apart from this difference the results are almost the same as those in the boltzmann limit .
the conductivities for large @xmath102 s are plotted in a different scale in the lower panel in fig.[fig_cond_short ] , compared with the high - energy expression of the boltzmann limit ( [ eq_sigma_weak_high ] ) and the approximate expression ( [ eq_sigma_scba ] ) .
the curves deviate from the boltzmann limit in the region @xmath130 and never fall below @xmath142 even in strong disorder .
the analytic expression becomes valid already for @xmath143 .
s , respectively . for the boltzmann conductivity in the lower panel we used the expression in ( [ eq_sigma_weak_high_long ] ) which is valid for @xmath121 . in the lower panel
the approximate result given by eq .
( [ eq_sigma_scba_long ] ) is also shown.,width=302 ]
for the long - range disorder , we consider scatterers with density @xmath144 and mean square amplitude @xmath145
. the two valleys @xmath6 and @xmath7 are now decoupled and the conductivity is written as the summation of their individual contributions .
the self - energy is given by @xmath146 which differs from ( [ eq_self0 ] ) for short - range disorder in that the prefactor is larger by a factor of 2 and that the summation is taken only in one valley .
however those two elements cancel and the self - energy becomes identical with ( [ eq_self ] ) in this notation .
we can omit the valley index @xmath88 completely .
the conductivity without the vertex correction , denoted by @xmath147 , is equivalent to ( [ eq_sigma ] ) .
the vertex part ( [ eq_vertex ] ) is given by @xmath148 where @xmath149 and @xmath57 are dimensionless quantities defined by @xmath150\nonumber\\ \noalign{\vspace{-0.250 cm } } & & \hspace{20mm}\times g_{j\vec{k}s}(\vare)g_{j\vec{k}s'}(\vare'),\end{aligned}\ ] ] with @xmath151 .
the conductivity correction including the valley degeneracy is then written as @xmath152 where @xmath153 and @xmath154 . in the boltzmann limit ,
only the term including @xmath155 and @xmath156 is relevant .
a straightforward calculation gives @xmath157 we can easily show that @xmath158 to obtain @xmath159 in the limit @xmath121 . in the case @xmath160 ,
on the other hand , @xmath161 and therefore @xmath162 .
the total conductivity combined with @xmath147 in ( [ eq_sigma_weak_high ] ) or ( [ eq_sigma_weak_low ] ) becomes @xmath163 in the case of large disorder @xmath164 , we can derive the analytic expression similarly to the short - range case . by using the self - energy ( [ eq_self_scba ] )
we can show that @xmath165 , @xmath166 , and @xmath167 , giving the vertex correction @xmath168 independent of energy .
the total conductivity including @xmath147 in ( [ eq_sigma_scba ] ) is written as @xmath169 + \frac{e^2}{\pi^2\hbar}\frac{2}{w } .
\label{eq_sigma_scba_long}\ ] ] for @xmath170 , this reduces to @xmath171 the conductivity is given by a universal value as in the short - range case in the limit of large disorder .
we show in fig .
[ fig_cond_long ] the conductivity numerically computed for the several @xmath102 s . the difference from the result in the short - range case is appreciable only in the clean limit @xmath172 , because the vertex correction gives only a shift of the order of @xmath173 .
the conductivity again approaches @xmath142 in the region @xmath130 in strong disorder .
the approximate result ( [ eq_sigma_scba0_long ] ) is valid for @xmath174 .
in secs . iv and v we have seen that the conductivity in the case of strong disorder @xmath164 deviates greatly from the boltzmann conductivity and converges to the order of @xmath1 .
this can be understood in terms of einstein s relation @xmath175 with the density of states @xmath176 and the diffusion constant @xmath177 .
the diffusion constant is written as @xmath178 , where @xmath179 is the average of the squared velocity over states at the fermi energy and @xmath101 is a relaxation time related to @xmath180 through the uncertainty relation @xmath181 .
if we neglect the @xmath0-linear term in the dispersion , we have @xmath182 and therefore @xmath183 with @xmath184 .
this is nothing but the boltzmann conductivity for @xmath185 . in the energy range @xmath186 , however , we have @xmath187 because we take an average over the states @xmath188 .
thus , the diffusion constant becomes @xmath189 independent of energy using @xmath190 . upon using @xmath191 ,
the conductivity becomes @xmath192 .
this conductivity is universal and independent of the band parameters and the strength of scattering .
it is also independent of short- or long - range disorder .
the situation becomes different in the weak - disorder limit @xmath139 , where @xmath193 at zero energy drops from the boltzmann conductivity almost by the factor @xmath102 .
this behavior is essentially equivalent to that in monolayer graphene , where the conductivity drops from the boltzmann to the universal value @xmath194 in the vicinity of the zero energy and the near - singular drop was ascribed to a reduction of the effective density of states contributing to the conductivity.@xcite the universal value was shown to be unaffected by magnetic fields,@xcite and similar near - singular behavior was shown to be present in various transport quantities.@xcite a universal conductivity at zero energy was also reported in the square tight - binding lattice model with one - half flux,@xcite which has @xmath0-linear dispersion as well .
this zero - energy conductivity can be explicitly estimated from ( [ eq_sigma ] ) with the density of states ( [ eq_rho_low ] ) and the square velocity ( [ eq_vsq_low ] ) . the integral turns out to be independent of the imaginary part of the self - energy and returns a universal value @xmath195 which is six times as large as the conductivity in the monolayer .
the extra factor comes from the product of the density of states and the square velocity , @xmath196 , which is larger than in the monolayer by a factor of 6 due to the existence of the satellite elliptic fermi pockets .
recently transport properties of bilayer graphene were studied experimentally.@xcite the resistivity exhibits a prominent peak at @xmath197 and decreases rapidly with the increase of the energy or the electron concentration .
this dependence is explained by eq .
( [ eq_sigma_scba ] ) or ( [ eq_sigma_scba_long ] ) ( or the lower panel of fig.[fig_cond_short ] or [ fig_cond_long ] ) qualitatively quite well with the disorder parameter @xmath198 , in the region @xmath199 @xmath31 . in particular , the observed peak resistivity @xmath200 k@xmath20 corresponds well to the present result for @xmath201 in the case of long - range disorder ( dominant usually ) and @xmath202 in the case of short - range disorder .
however , the observed resistivity seems to decrease much faster than that given by eq.([eq_sigma_scba ] ) or ( [ eq_sigma_scba_long ] ) with a constant @xmath102 for larger electron concentrations @xmath203 @xmath31 .
this might suggest that the effective range of the scattering potential can be comparable to the the electron wavelength at these electron concentrations .
we have studied the quantum transport in bilayer graphene in zero magnetic field with the self - consistent born approximation .
the coexistence of the @xmath0-linear and @xmath0-square dispersion is observed as two different behaviors in the boltzmann conductivity , like that in the usual two - dimensional metal and in a monolayer graphene with a massless dirac spectrum , which are energetically separated .
the conductivity in the scba deviates from the boltzmann limit in the case of strong disorder @xmath204 , and converges to @xmath205 ( per spin ) .
the conductivity in the long - range scatterers exhibits a qualitatively similar behavior to that in the short - range case , while the vertex correction gives a positive shift of the order of @xmath206 .
this work has been supported in part by the 21st century coe program at tokyo tech nanometer - scale quantum physics and by grants - in - aid for scientific research from the ministry of education , culture , sports , science and technology , japan .
numerical calculations were performed in part using the facilities of the supercomputer center , institute for solid state physics , university of tokyo .
99 k. s. novoselov , a. k. geim , s. v. morozov , d. jiang , y. zhang , s. v. dubonos , i. v. grigorieva , and a. a. firsov , science * 306 * , 666 ( 2004 ) . c. berger , z. song , t. li , x. li , a. y. ogbazghi , r. feng , z. dai , a. n. marchenkov , e. h. conrad , p. n. first , and w. a. de heer , j. phys .
b * 108 * , 19912 ( 2004 ) .
y. zhang , j. p.
small , w. v. pontius , and p. kim , appl .
. lett . * 86 * , 073104 ( 2005 ) .
k. s. novoselov , a. k. geim , s. v. morozov , d. jiang , m. i. katsnelson , i. v. grigorieva , s. v. dubonos , and a. a. firsov , nature * 438 * , 197 ( 2005 ) .
y. zhang , y. w. tan , h. l. stormer , and p. kim , nature * 438 * , 201 ( 2005 ) .
k. s. novoselov , e. mccann , s. v. morozov , v. i. falko , m. i. katsnelson , u. zeitler , d. jiang , f. schedin , and a. k. geim , nat . phys . * 2 * , 177 ( 2006 ) .
e. mccann and v. i. falko , phys .
lett . * 96 * , 086805 ( 2006 ) .
n. h. shon and t. ando , j. phys .
. jpn . * 67 * , 2421 ( 1998 ) .
j. w. mcclure , phys .
* 108 * , 612 ( 1957 ) .
j. c. slonczewski and p. r. weiss , phys
109 * , 272 ( 1958 ) .
w. w. toy , m. s. dresselhaus and g. dresselhaus , phys .
b * 15 * , 4077 ( 1977 ) .
a. misu , e. mendez , and m. s. dresselhaus , j. phys .
. jpn . * 47 * , 199 ( 1979 ) .
r. e. doezema , w. r. datars , h. schaber , and a. van schyndel , phys .
b * 19 * , 4224 ( 1979 ) .
h. ajiki and t. ando j. phys .
jpn . * 65 * , 505 ( 1996 ) .
charlier , x. gonze , and j. -p .
michenaud , phys .
b * 43 * , 4579 ( 1991 ) .
t. ando and t. nakanishi , j. phys .
jpn . * 67 * , 1704 ( 1998 ) .
t. ando , j. phys .
74 * , 777 ( 2005 ) . | the transport properties of a bilayer graphene are studied theoretically within a self - consistent born approximation .
the electronic spectrum is composed of @xmath0-linear dispersion in the low - energy region and @xmath0-square dispersion as in an ordinary two - dimensional metal at high energy , leading to a crossover between different behaviors in the conductivity on changing the fermi energy or disorder strengths .
we find that the conductivity approaches @xmath1 per spin in the strong - disorder regime , independently of the short- or long - range disorder . |
a number of candidates for a weyl semimetal have been previously proposed . however , existing proposals have proven difficult to carry out because they rely on magnetic ordering to break time - reversal symmetry or fine - tuning the chemical composition of an alloy .
magnetic order can be difficult to predict from first - principles , may be difficult to measure experimentally , and most importantly , may not form large enough domains in a real sample for the properties of the weyl semimetal to be preserved . fine - tuning chemical composition is typically challenging to achieve and introduces disorder , limiting the quality of single crystals .
for instance , a proposal for a weyl semimetal in y@xmath1ir@xmath1o@xmath25 @xcite assumes an all - in , all - out magnetic order , which is challenging to verify in experiment @xcite .
another proposed weyl semimetal , hgcr@xmath1se@xmath26 @xcite , has a clear ferromagnetic order , but because of the cubic structure there is no preferred magnetization axis , likely leading to the formation of many small ferromagnetic domains .
moreover , a proposal in hg@xmath27cd@xmath28mn@xmath29te @xcite requires straining the sample to break cubic symmetry , applying an external magnetic field and fine - tuning the chemical composition .
another proposal for the inversion breaking compounds labi@xmath30sb@xmath28te@xmath0 and labi@xmath30sb@xmath28te@xmath0 @xcite requires fine - tuning the chemical composition to within 5% . despite extensive experimental effort , a weyl semimetal has not been realized in any of these compounds .
we believe that in large part the difficulty stems from relying on magnetic order to break time - reversal symmetry and fine - tuning the chemical composition to achieve the desired band structure .
we note that , in contrast to time - reversal symmetry breaking systems , large single crystals of inversion symmetry breaking compounds exist , such as the large bulk rashba material bitei , where the crystal domains are sufficiently large so that the rashba band structure is clearly observed in photoemission spectroscopy @xcite .
we propose that taas overcomes the difficulties of previous candidates because it realizes a weyl semimetal in a stoichiometric , inversion - breaking , single - crystal material .
we have also studied other members of this class of compounds , tap , nbas and nbp .
they have the same crystal structure as taas .
our results show that their band structure is qualitatively the same as that of taas , in that the conduction and valence bands form line nodes in the absence of spin - orbit coupling and a gap is opened in the presence of spin - orbit coupling , leading to a weyl semimetal . in particular , since nb and p have quite weak spin - orbit coupling , our calculations suggest that nbp is better described as a topological nodal - line semimetal .
thus nbp might offer the possibility to realize novel nodal line band crossings in the bulk and `` drumhead '' surface states stretching across the nodal line on the surface @xcite .
next , we provide some general comments on the nature of the phase transition between a trivial insulator , a weyl semimetal and a topological insulator , illustrated in fig .
[ fig4]*a*. when a system with inversion symmetry undergoes a topological phase transition between a trivial insulator and a topological insulator , the band gap necessarily closes at a kramers point .
if we imagine moving the system through the phase transition by tuning a parameter @xmath31 , then there will be a critical point where the system is gapless . in a system which breaks inversion symmetry
, there will instead be a finite range of @xmath31 where the system remains gapless , giving rise to a weyl semimetal phase . in this way
, the weyl semimetal phase can be viewed as an intermediate phase between a trivial insulator and a topological insulator , where the bulk band gap of a trivial insulator closes and weyl points of opposite chiral charge nucleate from the bulk band touchings . as @xmath31
is varied , the weyl points thread surface states through the surface bz and eventually annihilate each other , allowing the bulk band gap to reopen with a complete set of surface states , giving rise to a topological insulator .
this understanding of a weyl semimetal as an intermediate phase between a trivial insulator and a topological insulator offers some insight into the closed fermi surfaces we find in taas around the @xmath32 point of the top surface and the @xmath33 point of the bottom surface .
we propose that these surface states reflect the topological invariant we would find if we annihilated the weyl points to produce a bulk insulator .
we consider the bottom surface , and annihilate the weyl points in pairs in the obvious way to remove all surface states along @xmath21 .
then , we can annihilate the remaining weyl points to produce two concentric fermi surfaces around the @xmath33 point . since this is an even number of surface states , we find that this way of annihilating the weyl points gives rise to a trivial insulator . if , instead , we annihilate the remaining weyl points to remove the fermi arc connecting them , only the closed fermi surface would be left around @xmath33 , giving rise to a topological insulator .
a similar analysis applies to the top surface . finally , we highlight an interesting phenomenon regarding how an electron travels along a constant energy contour in our predicted weyl semimetal , taas . fig .
[ fig4]*b * shows the calculated fermi surface contours on both the top and the bottom surfaces near the @xmath24 point .
we consider an electron initially occupying a state in one of the red fermi arcs , whose wavefunction is therefore localized on the top surface in real space .
we ask how the wavefunction evolves as the electron traces out the constant energy contour . as the electron moves along the fermi arc ( the red arc on the left - hand side of fig .
[ fig4]*c * ) , it will eventually reach a weyl point , where its wavefunction will unravel into the bulk .
the electron will then move in real space to the bottom surface of the sample .
then it will follow a fermi arc on that surface ( the blue arc on the left - hand side of fig .
[ fig4]*b * ) , and reach another weyl point .
the electron will next travel through the bulk again and reach the top surface , returning to the same fermi arc where it began .
this novel trajectory is predicted to show exotic surface transport behavior as proposed in ref .
this is one example of the unique transport phenomena predicted in weyl semimetals . to study these effects both from fundamental interest and for novel devices
, here we provides a promising candidate for a weyl semimetal in a stoichiometric , inversion - breaking , single - crystal material , as represented by taas .
we hope that our theoretical prediction will soon lead to the experimental discovery of the first weyl semimetal in nature .
* methods * first - principles calculations were performed by openmx code based on norm - conserving pseudopotentials generated with multi - reference energies and optimized pseudoatomic basis functions within the framework of the generalized gradient approximation ( gga ) of density functional theory ( dft ) @xcite .
spin - orbit coupling was incorporated through @xmath34-dependent pseudo - potentials @xcite . for each ta atom ,
three , two , two , and one optimized radial functions were allocated for the @xmath35 , @xmath36 , @xmath37 , and @xmath38 orbitals ( @xmath39 ) , respectively , with a cutoff radius of 7 bohr . for each as atom ,
@xmath40 was adopted with a cutoff radius of 9 bohr .
a regular mesh of 1000 ry in real space was used for the numerical integrations and for the solution of the poisson equation . a @xmath41 point mesh of ( @xmath42 ) for the conventional unit cell was used and experimental lattice parameters were adopted in the calculations .
symmetry - respecting wannier functions for the as @xmath36 and ta @xmath37 orbitals were constructed without performing the procedure for maximizing localization and a real - space tight - binding hamiltonian was obtained @xcite .
this wannier function based tight - binding model was used to obtain the surface states by constructing a slab with 80-atomic - layer thickness with ta on the top and as on the bottom .
turner , a. m. & vishwanath , a. beyond band insulators : topology of semi - metals and interacting phases .
preprint at http://arxiv:1301.0330 ( 2013 ) .
haldane , f. d. m. attachment of surface `` fermi arc '' to the bulk fermi surface : `` fermi - level plumbing '' in topological metals .
preprint at http://arxiv:1401.0529 ( 2014 ) .
balents , l. weyl electrons kiss .
_ physics _ * 4 * , 36 ( 2011 ) .
geim , a. k. & novoselov , k. s. the rise of graphene .
_ nature mater .
_ @xmath45 , 183 ( 2007 ) .
hasan , m. & kane , c. l. colloquium : topological insulators .
phys . _ * 82 * , 3045 ( 2010 ) .
qi , x. l. & zhang , s. c. topological insulators and superconductors .
mod . phys_. * 83 * , 1057 ( 2011 ) .
wang , z. _ et al . _
dirac semimetal and topological phase transitions in a@xmath0bi ( a = na , k , rb ) .
b _ @xmath46 , 195320 ( 2012 ) .
wang , z. _ et.al .
, _ three dimensional dirac semimetal and quantum spin hall effect in cd@xmath0as@xmath1 .
b _ @xmath47 , 125427 ( 2013 ) .
zyuzin , a. a. & burkov , a. a. topological response in weyl semimetals and the chiral anomaly .
b _ @xmath48 , 115133 ( 2012 ) .
liu , c .- x .
, ye , p. & qi , x .-
chiral gauge field and axial anomaly in a weyl semi - metal .
b _ @xmath49 , 235306 ( 2013 ) .
hosur , p. & qi , x .-
recent developments in transport phenomena in weyl semimetals .
_ comptes rendus physique _
@xmath50 , 857 ( 2013 ) .
parameswaran , s. a. , grover , t. , abanin , d. a. , pesin , d. a. & vishwanath a. probing the chiral anomaly with nonlocal transport in three - dimensional topological semimetals .
x _ @xmath51 , 031035 ( 2014 ) .
p. hosur , friedel oscillations due to fermi arcs in weyl semimetals .
b _ @xmath48 , 195102 ( 2012 ) .
ojanen , t. helical fermi arcs and surface states in time - reversal invariant weyl semimetals .
b _ @xmath49 , 245112 ( 2013 ) .
potter , a. c. , kimchi , i. & vishwanath , a. quantum oscillations from surface fermi - arcs in weyl and dirac semi - metals .
_ nature commun .
_ @xmath52 , 5161 ( 2014 ) .
wan , x. , turner , a. m. , vishwanath , a. & savrasov , s. y. topological semimetal and fermi - arc surface states in the electronic structure of pyrochlore iridates .
b _ * 83 * , 205101 ( 2011 ) .
burkov , a. a. & balents , l. weyl semimetal in a topological insulator multilayer .
_ * 107*,127205 ( 2011 ) .
xu , g. _ et al .
_ chern semi - metal and quantized anomalous hall effect in hgcr@xmath1se@xmath26 .
_ phys . rev .
* 107 * , 186806 ( 2011 ) .
bulmash , d. , liu , c .- x . &
prediction of a weyl semimetal in hg@xmath27cd@xmath28mn@xmath53te._phys .
b _ @xmath54 , 081106(r ) ( 2014 ) .
shapiro , m. c. _ et al . _ structure and magnetic properties of the pyrochlore iridate y@xmath1ir@xmath1o@xmath25 .
b _ @xmath46 , 214434 ( 2012 ) .
liu , h. _ et al . _ magnetic order , spin dynamics and transport properties of the pyrochlore iridate y@xmath1ir@xmath1o@xmath25 .
_ solid state commun .
_ @xmath57 , 1 ( 2014 ) .
ishizaka , k. _ et al .
_ giant rashba - type spin splitting in bulk bitei .
_ nature mater .
_ @xmath58 , 521 ( 2011 ) .
burkov , a. a. , hook , m. d. & balents , l. topological nodal semimetals .
b _ @xmath59 , 235126 ( 2011 ) .
perdew , j. p. burke , k. & ernzerhof , m. generalized gradient approximation made simple .
_ @xmath60 , 3865 ( 1996 . ) http://www.openmx-square.org/ , ozaki , t. variationally optimized atomic orbitals for large - scale electronic structures . _
b _ @xmath61 , 155108 ( 2003 ) .
theurich , g. & hill , n. a. self - consistent treatment of spin - orbit coupling in solids using relativistic fully separable _ ab _ _ initio _ pseudopotentials . _
b _ @xmath62 , 073106 ( 2001 ) .
weng , h. , ozaki , t. & terakura , k. revisiting magnetic coupling in transition - metal - benzene complexes with maximally localized wannier functions .
b _ @xmath63 , 235118 ( 2009 ) .
* weyl points and chiral charges in taas . * * a , * in the absence of spin - orbit coupling , there are two line nodes on the @xmath64 mirror plane , @xmath16 , and two line nodes on the @xmath65 mirror plane , @xmath17 . in the presence of spin - orbit coupling , each line node vaporizes into six weyl points .
the weyl points are denoted by small circles .
black and white show the opposite chiral charges of the weyl points .
* b , * the band structure around one of the red line nodes .
* c , * the band structure near one of the weyl points at @xmath66 .
the weyl point is shifted away from the @xmath64 mirror plane .
* d , * the projection of the bulk weyl points around the @xmath21 line . * e , * a schematic for the projection of all the weyl points on the ( 001 ) surface brillouin zone .
we denote the eight weyl points that are located on the @xmath67 plane as @xmath19 and the other sixteen weyl points as @xmath20 .
* f , * calculated spin texture @xmath68 in the vicinity of two weyl points with opposite chiral charges.,width=642 ] * schematic illustrations of several important aspects of an inversion breaking weyl semimetal . * * a , * a weyl semimetal can be understood as an intermediate phase between a trivial insulator and a topological insulator .
consider the evolution of weyl points and surface states as a function of a tuning parameter @xmath31 . in the trivial insulator phase
, there are no surface states . at a critical @xmath31 ,
the bulk conduction and valence bands touch each other . for higher values of @xmath31 ,
each bulk band touching splits into two weyl points with opposite chiral charges , and a fermi arc surface state connects them . in this regime ,
the system is a weyl semimetal . as @xmath31 increases further , eventually the weyl points pair up again , giving rise to another critical point .
for still higher values of @xmath31 , the weyl points annihilate and a full bulk band gap reopens , leaving a gapless surface state .
the system is then a topological insulator .
* b , * the calculated fermi surface of taas near the @xmath24 point . * c , * an electron tracing out a constant energy contour on a fermi arc will follow an unusual path in real space . initially , the electron wavefunction will be localized near , width=642 ] | the recent discoveries of dirac fermions in graphene and on the surface of topological insulators have ignited worldwide interest in physics and materials science .
a weyl semimetal is an unusual crystal where electrons also behave as massless quasi - particles but interestingly they are not dirac fermions .
these massless particles , weyl fermions , were originally considered in massless quantum electrodynamics but have not been observed as a fundamental particle in nature .
a weyl semimetal provides a condensed matter realization of weyl fermions , leading to unique transport properties with novel device applications .
such a semimetal is also a topologically non - trivial metallic phase of matter extending the classification of topological phases beyond insulators .
the signature of a weyl semimetal in real materials is the existence of unusual fermi arc surface states , which can be viewed as half of a surface dirac cone in a topological insulator . here
, we identify the first weyl semimetal in a class of stoichiometric materials , which break crystalline inversion symmetry , including taas , tap , nbas and nbp .
our first - principles calculations on taas reveal the spin - polarized weyl cones and fermi arc surface states in this compound .
we also observe pairs of weyl points with the same chiral charge which project onto the same point in the surface brillouin zone , giving rise to multiple fermi arcs connecting to a given weyl point .
our results show that taas is the first topological semimetal identified which does not depend on fine - tuning of chemical composition or magnetic order , greatly facilitating an exploration of weyl physics in real materials .
( note added : this theoretical prediction of november 2014 ( see paper in nature communications at http://www.nature.com/ncomms/2015/150612/ncomms8373/full/ncomms8373.html ) was the basis for the first experimental discovery of weyl fermions and topological fermi arcs in taas recently published in science ( 2015 ) at http://www.sciencemag.org/content/early/2015/07/15/science.aaa9297.abstract ) the rich correspondence between high energy and condensed matter physics has led to a deeper understanding of spontaneous symmetry breaking , phase transitions , renormalization and many other fundamental phenomena in nature , with important consequences for practical applications using magnets , superconductors and other novel materials @xcite .
recently , there has been considerable progress in realizing particles previously considered in high energy physics as emergent quasiparticle excitations of crystalline solids , such as majorana fermions and dirac fermions @xcite .
materials that host these exotic particles exhibit unique properties and hold promise for applications such as fault - tolerant quantum computation , low - power electronics and spintronics .
weyl fermions were originally considered in massless quantum electrodynamics , but have not been observed as a fundamental particle in nature @xcite .
recently , it was theoretically understood that weyl fermions can arise in some novel semimetals with nontrivial topology @xcite .
a weyl semimetal has an electron band structure with singly - degenerate bands that have bulk band crossings , weyl points , with a linear dispersion relation in all three momentum space directions moving away from the weyl point .
these materials can be viewed as an exotic spin - polarized , three - dimensional version of graphene . however , unlike the two - dimensional dirac cones in graphene @xcite , the three - dimensional dirac cones in na@xmath0bi and cd@xmath0as@xmath1 @xcite or the two - dimensional dirac cone surface states of bi@xmath1se@xmath0 @xcite , the degeneracy associated with a weyl point depends only on the translation symmetry of the crystal lattice .
this makes the unique properties associated with this electron band structure more robust .
moreover , due to its nontrivial topology , a weyl semimetal can exhibit novel fermi arc surface states .
previously , fermi arcs were only found in high-@xmath2 superconductors due to strong electron correlation effects .
recent theoretical advances in topological physics have now made it possible to realize fermi arc surface states in a weakly interacting semimetal .
both the weyl fermions in the bulk and the fermi arc states on the surface of a weyl semimetal are predicted to show unusual transport phenomena , which can be used in future device applications .
weyl fermions in the bulk can give rise to negative magnetoresistance , the quantum anomalous hall effect , non - local transport and local non - conservation of ordinary current @xcite .
fermi arc states on the surface are predicted to show novel quantum oscillations in magneto - transport and quantum interference effects in tunneling spectroscopy @xcite . because of the fundamental and practical interest in weyl semimetals , it is crucial that robust candidate materials be found .
it is theoretically known that a weyl semimetal can only arise in a crystal where time - reversal symmetry or inversion symmetry is broken . a number of magnetically ordered or inversion symmetry breaking materials have been proposed as weyl semimetal candidates @xcite . however , despite extensive effort in experiment , a weyl semimetal has yet to be realized in any of the compounds proposed thus far .
one concern is that the existing proposals require either magnetic ordering in sufficiently large domains @xcite or fine - tuning of the chemical composition to within 5% in an alloy @xcite , which are demanding in real experiments . here
we propose the first weyl semimetal in a stoichiometric , inversion - breaking , single - crystal material , taas .
unlike previous predictions , our proposal does not depend on magnetic ordering over sufficiently large domains , because our material relies on inversion symmetry breaking rather than time - reversal symmetry breaking .
the compound we propose is also stoichiometric and does not depend on fine - tuning chemical composition in an alloy .
single crystals of taas have been grown @xcite .
we believe that this material is promising for experimental realization for the weyl semimetal phase in this compound .
tantalum arsenide , taas , crystalizes in a body - centered tetragonal lattice system ( fig .
[ fig1]*a * ) . the lattice constants are @xmath3 and @xmath4 , and the space group is i4@xmath5md ( # 109 , @xmath6 ) .
the crystal consists of interpenetrating ta and as sub - lattices , where the two sub - lattices are shifted by @xmath7 , @xmath8 .
there are two ta atoms and two as atoms in each primitive unit cell .
it is important to note that the system lacks a horizontal mirror plane and thus inversion symmetry .
this makes it possible to realize an inversion breaking weyl semimetal in taas .
we also note that the @xmath9 rotational symmetry is broken at the ( 001 ) surface because the system is only invariant under a four - fold rotation with a translation along the out - of - plane direction . the bulk and ( 001 )
surface brillouin zones ( bz ) are shown in fig .
[ fig1]*b * , where high symmetry points are also noted .
the ionic model would suggest that the ta and as atoms are in the @xmath10 and @xmath11 valence states , respectively .
this indicates that the lowest valence band arises from @xmath12 electrons in as and @xmath13 electrons in ta , whereas the lowest conduction band primarily consists of @xmath14 electrons in ta .
however , we may expect ta @xmath14 electrons to have a broad bandwidth because of the wide extent of the atomic orbitals .
this leads to strong hybridization with the as @xmath12 states , which may suggest that the conduction and valence bands are not entirely separated in energy and have a small overlap , giving rise to a semimetal . in fig .
[ fig1]*c * , we present the bulk band structure in the absence of spin - orbit coupling .
the conduction and valence bands cross each other along the @xmath15 trajectory , which further indicates that taas is a semimetal . in the presence of spin - orbit coupling ,
the band structure is fully gapped along the high symmetry directions considered in fig .
[ fig1]*d*. however , weyl points which are shifted away from the high symmetry lines arise after spin - orbit coupling is taken into account .
below , we consider the weyl points in the bulk bz .
we also note that the double degeneracy of bands is lifted in the presence of spin - orbit coupling except at the kramers points , which confirms that taas breaks inversion symmetry but respects time - reversal symmetry .
to better understand the weyl points in taas , we first examine the band crossings near the fermi level in the absence of spin - orbit coupling .
these band crossings take the form of a closed curve , or line node , on the @xmath16 and @xmath17 mirror planes of the bz , shown in red and blue in fig .
[ fig2]*a*. we present the band structure near one of the red nodal lines , on the @xmath16 mirror plane , in fig . [ fig2]*b*. the conduction and valence bands dip into each other , giving rise to a line node crossing .
next , we include spin - orbit coupling .
this causes each line node to vaporize into six weyl points shifted slightly away from the mirror planes , shown as small circles in fig .
[ fig2]*a*. there are 24 weyl points in total : 8 weyl points on the @xmath18 plane , which we call @xmath19 , and 16 weyl points away from the @xmath18 plane , which we call @xmath20 .
we present the band structure near one of the weyl points @xmath19 in fig .
[ fig2]*c * , where a point band touching is clearly observed .
next , we consider how the weyl points project on the ( 001 ) surface bz .
we show a small region in the surface bz around the @xmath21 line , which includes the projections of six weyl points , two from @xmath19 and four from @xmath20 .
a schematic of the projections of all 24 weyl points on the surface bz is shown in fig .
[ fig2]*e*. we find that for all points @xmath20 , two weyl points project on the same point in the surface bz . we indicate the number of weyl points that project on a given point in the surface bz in fig .
[ fig2]*e*.
lastly , we study the chirality of the weyl points by calculating the berry flux through a closed surface enclosing a weyl point .
we call positive the chiral charges which are a source of berry flux and negative the chiral charges which are a sink of berry flux .
we color the positive chiral charges white and the negative chiral charges black . in fig .
[ fig2]*f * , we show the spin texture of the band structure near two weyl points @xmath19 . in this case , the spin texture is found to be consistent with the texture of the berry curvature .
moreover , it turns out that the points @xmath20 project on the surface bz in pairs which carry the same chiral charge .
another key signature of a weyl semimetal is the presence of fermi arc surface states which connect the weyl points in pairs in the surface bz . we present calculations of the ( 001 ) surface states in fig . [ fig3 ] .
we show the surface states on the top surface in fig .
[ fig3]*a * and the bottom surface in fig .
[ fig3]*b*. we find surface fermi arcs that connect weyl points of opposite chirality in pairs . to better understand the rich structure of the fermi arcs , we show a schematic of the surface states on the top surface in fig .
[ fig3]*c * and the bottom surface in fig . [ fig3]*d*. note that one fermi arc connects each pair of points @xmath19 . however , two fermi arcs connect to each projection of points @xmath20 , because they project in pairs with the same chiral charge , as discussed above .
this leads to fermi arcs which connect the points @xmath20 in a closed loop of surface states .
the largest fermi arc loop on the top surface threads through four projected weyl points in the surface bz .
we also note that the fermi surface of the surface states from the top is very different from that of the bottom ( figs .
[ fig3]*a , b * ) , consistent with broken inversion symmetry in this system .
in addition , because the presence of a surface breaks the @xmath9 screw symmetry , the surface states are also very different along the @xmath22 and @xmath23 directions of the surface bz . finally , we observe closed fermi surfaces which do not intersect weyl points and do not form fermi arcs .
these extra fermi surfaces reflect how different ways of annihilating the weyl points would give rise to an insulator with a different topological invariant ( see the discussion below ) .
we present a particularly simple set of fermi arcs arising near the @xmath24 point in fig .
[ fig3]*e * , including surface states from both the top and bottom surfaces . to visualize the arc nature of the surface states , we present three energy dispersion cuts along the directions indicated in fig .
[ fig3]*e*. along cut 1 , shown in fig .
[ fig3]*f * , we see a dirac cone connecting the bulk valence and conduction bands across the bulk band gap , exactly like a topological insulator . along cut 2 , shown in fig . [ fig3]*g * , we see the projected bulk weyl cones , with surface states which pass through the weyl points .
lastly , in cut 3 , shown in fig .
[ fig3]*h * , we observe a full band gap .
the surface states along this cut are trivial because they do not connect across the bulk band gap . |
study of multidimensional and multi - particle objects is qualitatively more complicated than that of the one - dimensional case .
when there is no symmetry in the potentials of interaction between particles , these systems are described by partial differential equations with unseparable variables . just for this reason the inverse scattering problem in 2- and 3-dimensional spaces was formulated by faddeev @xcite and newton @xcite , novikov and henkin @xcite at a much later than the 1-dimensional problem .
the foundation for developing the 2-dimensional inverse problem in finite differences was laid by berezanskii @xcite who developed the theory of orthogonal polynomials for the jacobi infinite matrix .
the darboux transformations in quantum mechanics for the schrdinger differential and difference 1-dimensional equation have much in common with the spectral transformations for orthogonal polynomials @xcite .
it is therefore expedient to analyze the spectral inverse problem for the discrete 2-dimensional schrdinger equation on the basis of the technique of orthogonalization of polynomials .
the formulae of the inverse problem with degenerate kernels are closely related to bargmann - darboux transformations @xcite@xcite .
the one - dimensional darboux transformations has already found wide applications in quantum mechanics and in the theory of nonlinear integrable systems @xcite . in the present paper , a discrete version of the gelfand - levitan inverse spectral problem
is considered for the two - dimensional lattice schrdinger equation .
the two - dimensional finite - difference inverse problem , based on the procedure of orthogonalization of polynomial vectors , is a generalization of the one - dimensional procedure given in @xcite . on the basis of the obtained formulae of the inverse problem ,
the relations are derived for discrete bargmann - darboux transformations in two dimensions .
the suggested algebraic approach allows one to construct families of discrete potentials in an explicit form and the corresponding solutions .
consider the schrdinger equation whose hamiltonian is tridiagonal in a certain basis with respect to both the coordinate variables @xmath0 and @xmath1 @xmath2 the coefficients @xmath3 are assumed to be real and represent discrete potentials ; @xmath4 are discrete wave functions , @xmath5 is an integer point of the half - plane , @xmath6 , @xmath7 is a spectral parameter . the index @xmath0 can vary from @xmath8 to @xmath9 and the index @xmath1 can vary from @xmath10 to @xmath9 or from @xmath8 to @xmath9 ; when @xmath11 and @xmath12 , it is a special class of restricted problems .
the equation in finite differences ( [ 1.1 ] ) can be represented by the expression with operator coefficients @xmath13 when we treat one of the variables , @xmath0 , as the only discrete coordinate ; and the other , @xmath1 , as the channel index .
if we set @xmath14 , the action of the schrdinger discrete operator @xmath15 on the vector @xmath16 is represented by the action of the jacobi block matrix @xmath17 on @xmath18 @xmath19 whose elements @xmath20 and @xmath21 are matrices at each fixed @xmath0 , and each element @xmath22 of the vector @xmath18 corresponds to the vector @xmath23 , in the other space variable @xmath24 @xmath25 @xmath26 it is clear that the operator @xmath15 is hermitian . note that the coupling of equations ( [ 1.1 ] ) or ( [ 1.2 ] ) with respect to @xmath1 occurs only between immediate neighbours and for this reason the symmetric coupling matrix @xmath27 in ( [ 1.2 ] ) is tridiagonal and relates the functions @xmath4 at @xmath28 , unlike the conventional multichannel case . from the jacobi block matrix ( [ 1.j ] )
tridiagonal in the variable @xmath0 it is seen that every vector of the solutions @xmath22 is connected with the vectors of solutions @xmath29 at the neighboring points @xmath30 . as a result ,
if we take nonhomogeneous boundary conditions at one end of the interval @xmath31 , we can obtain the solutions on the whole interval by moving by subsequent steps from that end .
it turns out that these vectors of solutions as functions of the spectral parameter @xmath32 are polynomials of @xmath32 with matrix coefficients @xcite which can be orthogonalized in the spectral measure .
the spectral inverse problem is reduced to the construction of potential matrices @xmath20 , @xmath21 and an unknown system of orthonormal polynomials with the use of the known system of orthonormal polynomials corresponding to the finite - difference equation ( [ 1.2 ] ) but with the known matrices @xmath33 and @xmath34 .
introduce auxiliary solutions @xmath35 and @xmath36 to eq.([1.1 ] ) with the help of the boundary conditions @xmath37 where @xmath38 is a point on the @xmath1 axis .
the solutions @xmath39 satisfy the same equation ( [ 1.1 ] ) with the known potentials @xmath40
@xmath41 @xmath42 as a first boundary condition , we can change the zeroth condition to a more general one corresponding to the homogeneous boundary conditions @xmath43 . in what follows
we assume the step of finite - difference differentiation @xmath44 to equal 1 . according to the conditions ( [ 1.3 ] ) ,
the functions @xmath45 and @xmath36 are equal to zero on the lines @xmath46 , @xmath47 except for the points with @xmath48 on the line @xmath49 where @xmath50 . from these values of @xmath51 and @xmath52
one can find the functions @xmath45 and @xmath36 at all other points of the coordinate network by using the recurrence equation ( [ 1.1 ] ) . since the matrix equation ( [ 1.2 ] )
is tridiagonal in the variable @xmath0 , and owing to the boundary conditions ( [ 1.3 ] ) , the vectors @xmath53 of solutions
@xmath54 in the spectral variable @xmath7 are polynomials of the @xmath0th degree . at @xmath55
the vector @xmath56 is a polynomial of the first degree ; at @xmath57 , the vector @xmath58 is a polynomial in @xmath7 of the second degree , and so on .
it appears , however , that different elements @xmath59 of the vector of solutions @xmath60 have different degrees of polynomials . because of the three - point coupling with respect to @xmath1 the functions @xmath59 are polynomials in @xmath32 of the degree @xmath61 .
the maximum @xmath0th degree belongs to the elements with @xmath48 and the elements @xmath62 vanish at all @xmath38 lying out of the values @xmath63 .
the distribution of polynomial degrees is shown in fig.1 .
+ @xmath64 and similarly for @xmath65 @xmath66 the spectral matrix elements @xmath67 ( or @xmath68 ) are determined by the boundary values of the special solutions @xmath4 ( or @xmath69 obeying the zeroth condition @xmath70 @xmath71 .
for example , for @xmath72 states of the discrete spectrum , the matrix @xmath73 is the sum of @xmath72 terms formed from productions of the column vectors @xmath74 and the row vectors @xmath75 @xmath76 where the elements @xmath77 are defined by @xmath78 and @xmath79 is the heaviside step function equal to 1 at @xmath80 and to zero when @xmath81 .
a solution @xmath4 to eq.([1.1 ] ) with the zeroth condition ( or a more general homogeneous condition ) where both of them exist when one of the boundary conditions is the same .
] can be obtained by multiplying the matrix @xmath82 by the vector @xmath83 : @xmath84 at every fixed @xmath0 and @xmath1 the summation over @xmath38 is finite owing to @xmath85 outside the interval @xmath63 . using the procedure of orthogonalization of polynomials
we construct unknown polynomial solutions @xmath86 normalized with the spectral weight @xmath87 as a linear combination of the known polynomial solutions @xmath65 , orthogonal with respect to the measure @xmath88 @xmath89 as one can see above , the validity of this relation is a consequence of the fact that both the functions , @xmath65 and @xmath86 , obey the same discrete schrdinger equation ( [ 1.1 ] ) and the same boundary conditions ( [ 1.3 ] ) , given on the lines @xmath90 and @xmath49 .
therefore , the solutions @xmath86 and @xmath65 are polynomials in @xmath7 of the same degree @xmath61 , but with different coefficients .
the matrix of the polynomial solutions @xmath82 , which is of degree @xmath0 , is orthogonal to every polynomial matrix of degree lower than @xmath0 and hence to every @xmath91 for @xmath92 @xmath93 equation ( [ 1.4 ] ) is a discrete analog of the volterra integral equation , in which the coefficients @xmath94 are determined by the condition of orthogonality of the vector - functions @xmath95 orthogonal with the spectral measure @xmath73 to the functions @xmath91 orthogonal with the weight matrix @xmath96 , when @xmath97 @xmath98 the volterra equations ( [ 1.4 ] ) have a triangular form , @xmath99 for @xmath100 .
it is easy to see from ( [ 1.4 ] ) and orthogonality of @xmath101 ( [ 1.7 ] ) that for @xmath102 @xmath103 inserting ( [ 1.4 ] ) into ( [ 1.6a ] ) for @xmath104 we obtain the following system of equations for the orthogonalization coefficients @xmath94 @xmath105 where @xmath106 equation ( [ gl ] ) is a two - dimensional analog of the gelfand - levitan integral equations for finite - difference equation ( [ 1.1 ] ) . however ,
this system is not sufficient to determine the coefficients @xmath107 .
a supplementary system of equations is obtained upon substituting ( [ 1.4 ] ) into the complete relation ( [ 1.6 ] ) at @xmath108 and @xmath109 @xmath110 now let us find connections between the potential coefficients and orthogonalization coefficients @xmath94 .
to this end , we insert expression ( [ 1.4 ] ) for the polynomial functions @xmath86 in terms of @xmath65 into the schrdinger difference equation ( [ 1.1 ] ) @xmath111 we transform the r.h.s . of the above relation by means of the substitution of @xmath112 from the finite difference equation ( [ 1.1a ] ) for @xmath65 with the known potentials
@xmath113 , @xmath114 @xmath115 further , we take advantage of the orthogonality relation ( [ 1.7 ] ) for the matrix functions @xmath116 orthogonal with the weight matrix @xmath96 . multiplying expression ( [ 1.8 ] ) with its transformed r.h.s .
( [ 1.8a ] ) by @xmath117 , integrating over @xmath118 , and summing up over the indices @xmath38 and @xmath119 , we arrive at the relationship between the potentials @xmath120 and the coefficients @xmath121 @xmath122 the relations for the coefficients @xmath123 and @xmath124 are established in a similar manner . to determine @xmath125 , eq.([1.8 ] )
is multiplied by @xmath126 with ( [ 1.8a ] ) taken into account and integrated with the weight @xmath88 by using the orthogonality ( [ 1.7 ] ) . as a result
, we have @xmath127 the relation for @xmath128 is derived analogously , only ( [ 1.8 ] ) is multiplied by @xmath129 @xmath130 substituting ( [ 1.9 ] ) into ( [ 1.12 ] ) we arrive at @xmath131 at @xmath132 , @xmath133 , the derived generalized expressions turn into more simple ones presented in @xcite .
the two - dimensional finite - difference inverse problem under consideration is also a generalization of that @xcite with the potential coefficients @xmath134 , @xmath135 , connected nevertheless in a special way . in principle
, it is easy to formulate the problem of restoring the matrix @xmath136 in ( [ 1.2 ] ) with all nonzeroth elements , like in a multichannel problem @xcite .
the latter would correspond to the potential being nonlocal with respect to one of the coordinate variables ( in our case @xmath24 ) .
if the consideration were made in the polar coordinate system , nonlocality with respect to discrete angles would occur . in the case of continuous coordinates ,
the inverse problem for the potential , nonlocal relative to angles , was considered by kay and moses @xcite .
in this section , we describe the algebraic procedure by taking into considaration simple kernels @xmath137 in the form of a sum of several terms with a factorized coordinate dependence @xmath138 here the functions @xmath139 are combined as elements of the vector @xmath140 obtained as a product of the matrix solutions @xmath141 taken at eigenvalues @xmath142 of the reconstructed @xmath15 and the vector @xmath143 @xmath144 the elements @xmath145 form the normalization matrix @xmath146 with elements @xmath147 corresponding to the bound state @xmath148 , @xmath149 . like @xmath137 , the ortogonalization kernel @xmath121 is also presented as a sum of several factorized terms .
really , substituting ( [ 1.q ] ) into the gelfand levitan equation ( [ gl ] ) , we obtain @xmath150 noting that the expression in braces is the solution @xmath151 ( [ 1.4 ] ) at @xmath152 of eq.([1.1 ] ) with the desired potentials @xmath153 and
@xmath123 , we immediately receive @xmath154 it is evident now that the new wave functions @xmath86 , determined by ( [ 1.4 ] ) with the kernel @xmath107 taken in the form ( [ 1.k ] ) , are related to the old ones @xmath65 by @xmath155 in view of ( [ 1.k ] ) for @xmath94 in eqs .
( [ 1.9 ] ) , ( [ 1.11 ] ) and ( [ 1.12 ] ) , one can immediately write expressions for discrete potentials in the closed form @xmath156 @xmath157 and @xmath158 the solutions @xmath159 have to be found from the gelfand - levitan equations ( [ gl ] ) and ( [ 1.7a ] ) taking account of ( [ 1.q ] ) and ( [ 1.k ] ) .
thus , the operator @xmath94 defined by ( [ 1.k ] ) transforms the solutions @xmath160 of eq .
( [ 1.1a ] ) into the solutions @xmath161 of eq.([1.1 ] ) , determined by ( [ 1.4 ] ) or ( [ 1.4bar ] ) , with the potentials @xmath153 and @xmath123 defined by ( [ a - nm ] ) , ( [ b - nm ] ) and ( [ c - nm ] ) .
it is not difficult to see from the definitions ( [ q ] ) and ( [ 1.6a ] ) that the kernels @xmath137 and @xmath107 like ( [ 1.q ] ) and ( [ 1.k ] ) can be obtained provided that the spectral weight functions @xmath73 and @xmath162 for both the sets of potentials coincide except , for instance , @xmath72 eigenvalues at @xmath142 .
this permits one to construct potentials with @xmath72 new bound states by using ( [ a - nm ] ) , ( [ c - nm ] ) and ( [ b - nm ] ) or generate the family of spectral - equivalent potentials whose spectra coincide @xmath163 and it is only the normalization factors @xmath164 that are different . in the latter case @xmath137 is taken in the form @xmath165 and the above procedure can be used to construct spectral - equivalent operators @xmath166 and @xmath15 . in spite of a complicated form of the expressions for the potentials ( [ a - nm ] ) , ( [ b - nm ] ) and ( [ c - nm ] ) , they are simplified for a large set of particular cases .
for example , if we deal with the free discrete schrdinger equation as a reference one , fixed by the choice @xmath167 , @xmath168 .
let us now consider another simple case when the hamiltonians differ only by spectral data at one bound state . in this case , the summation over @xmath169 in all formulae ( [ 1.4bar ] ) ( [ c - nm ] ) vanishes .
general solutions @xmath170 from ( [ 1.4bar ] ) at arbitrary @xmath7 can be written as @xmath171 where @xmath172 is a special solution of the schrdinger equation ( [ 1.1a ] ) for the value of the spectral parameter @xmath173 and @xmath174 is the solution of eq.([1.1 ] ) with the same eigenvalue @xmath173 .
new potentials are expressed in terms of the known old potentials @xmath175 , @xmath176 , and @xmath177 , functions @xmath178 and functions @xmath4 that can be determined from the second gelfand - levitan equation ( [ 1.7a ] ) @xmath179 @xmath180 and @xmath181 it should be noted that relationships between potentials and functions can be obtained within the darboux transformation method or factorised method without using formulae of the inverse problem . _
connection between darboux transformations and inverse problem ones .
_ it is interesting to note that transformation ( [ 1.4b ] ) with one bound state corresponds to darboux transformation for the finite - difference equation ( [ 1.1 ] ) .
indeed , let us search for a solution @xmath86 of eq.([1.1 ] ) with some initially unknown potentials in the form ( [ 1.4b ] ) .
next it is necessary to find conditions for the potentials @xmath153 and @xmath123 and special functions @xmath4 at which general solutions @xmath182 specified by ( [ 1.4b ] ) will satisfy the discrete schrdinger equation ( [ 1.1 ] ) .
substitute ( [ 1.4b ] ) into ( [ 1.1 ] ) @xmath183 transform the r.h.s . of ( [ cond ] ) substituting @xmath184 from ( [ 1.1a ] ) @xmath185.\end{aligned}\ ] ]
further , to obtain the relations for potentials @xmath186 , @xmath187 and @xmath123 , multiply eq.([cond ] ) with its transformed r.h.s .
( [ rhs ] ) by @xmath188 , @xmath189 and @xmath190 and take into considaration the completeness relation ( [ 1.7 ] ) for the functions @xmath191 .
the expressions for @xmath192 , @xmath124 and @xmath123 thus derived coincide with formulae ( [ a - nm1 ] ) , ( [ b - nm1 ] ) and ( [ c - nm1 ] ) , correspondingly , obtained from the formulae of the inverse problem .
the gelfand - levitan spectral inverse problem for the discrete two - dimensional schrdinger equation is considered on the basis of the berezanskii technique of orthogonalization of polynomial matrices . by using the derived formulae of the inverse problem , discrete bargmann
darboux transformations in two dimensions are given .
analytic relationships are established between the solutions for two different sets of discrete potentials and the potentials themselves .
+ i thank dr.v.m.muzafarov and e.p.velicheva for helpful discussions . + a part of this work was done in june of 1998 at the erwin schrdinger international institute for mathematical physics , wien , austria .
faddeev l.d .
1974 _ sov.probl.math . _ * 3 * , 93 - 180 .
newton r.g .
1980 _ j.math.phys . _
* 21 * 1698 ; 1981 * 22 * , 631 ; ibid , 2191 .
newton r. 1989 _ inverse schrdinger scattering in three dimensions _
( springer - verlag .
berlin heidelberg / new york ) .
matveev v.b . and
salle m.a . 1979 _ lett.matt.phys . _ * 3 * 425 .
matveev v.b . and
salle m.a .
1991 _ darboux transformations and solitons _ springer series in nonlinear dynamics ( springer - verlag , berlin heidelberg / new york ) ; spiridonov v. 1997 _ j.phys.a , math.gen._ * 30 * l15-l20 .
case k.m . and kac m. , 1973 _ j.math.phys_. , * 14 * 594 - 603 . | a discrete version of the two - dimensional inverse scattering problem is considered . on this basis ,
algebraic transformations for the two - dimensional finite - difference schrdinger equation are elaborated .
generalization of the technique of one - dimensional darboux transformations for the two - dimensional finite - difference schrdinger equation is presented .
ideas of bargmann - darboux transformations for the differential one - dimensional multichannel schrdinger equation are used for the two - dimensional lattice schrdinger equation .
analytic relationships are established between different discrete potentials and the corresponding solutions .
inverse problem and darboux transformations for two - dimensional finite - difference schrdinger equation a.a .
suzko joint institute for nuclear research , dubna |
alfvn waves have been invoked as a driving force behind solar coronal heating and wind acceleration since their inception @xcite .
while alfvnic waves have been detected in situ by several spacecraft in and out of the ecliptic @xcite , in the solar atmosphere their presence had only been inferred through the large non - thermal broadening of coronal emission lines ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
only recently were they directly observed and measured in the chromosphere @xcite with the solar optical telescope ( sot ; * ? ? ?
* ) on the _ hinode _ spacecraft @xcite and in the corona with the coronal multi - channel polarimeter ( comp ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a subsequent investigation with the _ solar dynamics observatory _ atmospheric imaging assembly ( aia ; * ? ? ? * ) confirmed the ubiquity of coronal wave motions @xcite .
unfortunately , there is a significant discrepancy in the energy estimates between these three direct observations .
the alfvnic wave energy flux in doppler motions of the 10747 emission line observed by comp is @xmath00.01 wm@xmath1 @xcite , some four orders of magnitude less than the 100 wm@xmath1 required to balance the radiative losses of the quiet corona @xcite and/or accelerate the fast solar wind @xcite .
however , the observations of sot indicate that an alfvnic wave energy flux of order 4 kwm@xmath1 is present in the chromosphere @xcite and that even a very low transmission ( @xmath03% ) of these waves into the corona ( due to reflection off the transition region , * ? ? ?
* ) will result in an energy flux of the order of @xmath0100 wm@xmath1 into the corona or solar wind .
in addition , the presence of _ at least _ that much wave energy is readily observed in the quiescent ( quiet and coronal hole ) corona with sdo / aia @xcite .
this leaves us in a quandary : where is the missing wave energy in the comp observations ?
interestingly , the coronal wave energy flux inferred from the sot observations ( and measured by aia ) is consistent with the amplitude of non - thermal broadening observed by comp , as reported by @xcite .
so , perhaps the real energy flux of alfvnic waves in the corona is hidden from comp . in this scenario
, the vastly different spatial scales of the instruments involved would result in considerable superposition of low frequency motions resolved by sot and aia but unresolved by comp , quenching the resolved velocity amplitude , energy flux , and increasing the non - thermal line broadening of the observed line profiles in the optically thin corona .
an alternative scenario faces the challenging task of simultaneously explaining the large non - thermal broadening in the corona , and the observational evidence from aia of strong alfvnic waves throughout the corona .
perhaps the filling factor of the waves observed with aia is much lower than estimated and/or the bulk of the alfvnic waves ( coming up from the chromosphere ) are actually damped and/or dissipated in the first few thousand kilometers above the limb , leaving only low resolved wave amplitudes for comp ?
perhaps the non - thermal broadening of the coronal emission line is then caused by unresolved flows into and out of the plane - of - the - sky ( pos ) ? in this letter we investigate these possible scenarios , address the apparent mis - match of alfvnic wave energies observed by sot , aia , and comp and thus constrain the potential sources of non - thermal broadening , the invariant hidden , or `` dark '' energy , in the quiet corona .
we expand upon a simple forward model @xcite that explained the center - to - limb variation of transition region lines by taking into account that sot and aia observed a lower atmosphere dominated by a myriad of extremely fine - scale , field - aligned and independently moving structures , called spicules , with spatial widths @xmath2300 km ( considerably smaller than the @xmath04,500 km comp spatial sampling ) that exhibit significant transverse motions @xcite . in the following section we discuss the observations that drive the model . in sect .
method , we discuss how the sot and aia observations are used to drive the model . in sect .
analysis we look at the properties of the model that can reproduce the observed comp behavior .
the comp observations used in this paper were obtained on december 26 , 2011 from 20:16 to 21:45ut in three wavelengths ( 10745.2 , 10746.5 , and 10847.8 ) across the 10747 line , with an exposure time of 250ms at each position .
the image groups were obtained at a cadence of 30 seconds . from the three wavelength positions we measure the peak line intensity , doppler shift , and linewidth of the 10747 line .
fitting a histogram to the measured velocities allows us to determine the coronal rotation rate and estimate the `` rest '' wavelength to be 10746.08 ( assuming that half of the field of view is red and half is blue - shifted ) . for an instrumental width ( @xmath3 ) of 21km / s and a thermal width ( @xmath4 ) of 21km / s for ( using @xmath5 , for an ion of mass @xmath6 and peak formation temperature of @xmath7 assuming that ion and electron temperatures are equal ) we determine the non - thermal width @xmath8 @xmath9 , where @xmath10 is the 1/e width of the fitted line profile . the top row of fig .
f1 compares an aia 193 image ( panel a ) taken at the start of the timeseries with the comp @xmath8 ( panel b ) and root - mean - square of the doppler velocity during the timeseries ( @xmath11 ; panel c ) .
the latter is dominated by the resolved , low - frequency , and low - amplitude alfvnic waves that are seen to propagate in the corona .
panel d shows a two - dimensional histogram of @xmath11 and @xmath8 .
we find a strong correlation between @xmath11 and @xmath8 .
this is an indication that the two measures are intimately tied , i.e. , the variation of the resolved wave amplitude during the timeseries and the non - thermal broadening of the line in a pixel / voxel are co - dependent .
this suggests that whatever causes the strong non - thermal line broadening is related to the alfvnic waves that comp observes .
note the clear cut - off at @xmath8=17 - 18km / s and the enclosing `` correlation wedge '' of the two dot - dashed lines .
this wedge is a robust feature of the comp observations , shifting slightly left and right and tilting slightly up and down from day to day as the physical structure of the corona above the limb changes .
supporting movies of sdo / aia ( ) 193 , comp i , v , and @xmath10 can be found in the electronic edition of the journal , the latter clearly illustrate the ubiquity of the transverse doppler motions .
hinode / sot observations have shown that the chromospheric limb is dominated by superposition of a large number of long , thin ( @xmath12200 km ) spicules that extend to heights of 5 - 15 and that undergo significant motions along and perpendicular to the line - of - sight ( los ) .
@xcite demonstrated that there are two types of spicules present in the _ hinode _ limb observations .
type i spicules evolve on timescales of 3 - 5 minutes reaching velocities of 10 - 40km / s along their long axis @xcite .
type ii spicules on the other hand , occur on timescales of 10 - 120s and reach apparent longitudinal velocities of 50 - 150km / s .
@xcite have shown , using a statistical approach , that these chromospheric features undergo vigorous alfvnic motions with amplitudes around @xmath13km / s and periods ranging from 100 to 500s .
@xcite , using a similar approach , demonstrated that an energetically relevant portion of the alfvnic energy observed in the chromosphere by sot is indeed visible in emission formed at coronal temperatures .
those waves have made it `` into '' the corona - these observations were made possible by the high s / n , low scattered light , high spatial resolution observations of aia .
the imaged wave motions have amplitudes , periods , and speeds commensurate with those observed by sot near the limb , and can be considered the coronal extension @xcite of the chromospheric features studied by @xcite . the observed phase speeds of these coronal alfvnic waves reach 500 - 1000km / s , values that are commensurate with those observed by comp @xcite .
we use the high resolution sot @xcite and aia @xcite observations as a guide for the transverse motions that occur on a spatial scale that is far smaller than comp resolution .
we develop a ( simple ) monte - carlo forward model of alfvnic waves in the corona based on the concept of a `` thread '' , or elementary oscillating structure .
our simple picture is that any pixel in the optically thin model corona can be populated with @xmath14 independent threads and the 10747 signal observed is given ( at any time ) by summing the emission of all of the threads in any given pixel .
the properties of any single thread are as follows : a thread can appear at any time with a lifetime chosen from a gaussian distribution of 100 @xmath15 20s ; has a uniform brightness ; and is subject to alfvnic motion governed by a randomly chosen period , amplitude and phase .
here we consider swaying motions , since those are clearly observed with aia .
we assume a polarization angle chosen from a uniform distribution between 0 and @xmath16 with respect to the los , and apply a @xmath17 correction factor to accommodate the fact that not all threads are radial ( where @xmath18 follows a gaussian distribution of 0@xmath1520 ) .
the sinusoidal alfvnic motions of the thread have a period drawn from a distribution based on the comp power spectrum ( see , e.g. , fig . 2 of * ? ? ?
* ) , a phase from a uniform distribution between 0 and @xmath16 and an amplitude from a gaussian distribution around x@xmath155 km / s ( where x can be any value @xmath195km / s ) .
each thread emits a gaussian line profile of width 21km / s ( equivalent to @xmath4 of the plasma ) at the given los velocity determined from its period , amplitude and phase .
the resulting line profile in a pixel is the sum of the gaussians of the @xmath14 individual threads .
( a ) and @xmath20 ( b ) measured from the monte - carlo simulated timeseries of three - point 10747 measurements for a range of thread numbers and input wave amplitudes .
the white ( dot - dashed and solid ) lines show the 15 and 45km / s values of @xmath8 respectively , while the black ( dot - dashed and solid ) lines correspond to the 0.5 and 1.5km / s values of @xmath11 respectively .
[ f3],width=321 ] we then observe these synthetic profiles in the same fashion as the comp observations @xcite by sampling the 10747 line at the three wavelengths above with a gaussian filter width of 1.3 , and fitting the ensemble line profiles with a single gaussian to derive i , v , and @xmath20 .
figure f2 shows sample ( pixel ) timeseries of the doppler velocity ( and @xmath11 ; left column ) and non - thermal linewidth ( @xmath8 ; right column ) for a range of thread numbers ( bottom to top : 200 , 500 , 1000 , and 2000 ) and a fixed 40km / s wave ( peak - to - peak ) amplitude ( @xmath21 ) .
notice the diminishing values of @xmath11 with increasing thread number - these values eventually saturate at the gaussian fitting accuracy of the three - point measurements ( @xmath00.1 of the spectral resolution ) .
notice also that @xmath8 @xmath22@xmath21 and the gradual increase in @xmath8 with increasing thread number and the reduction of its variance with time - recalling that @xcite noticed an invariance of the linewidths with time in their original measurements . to investigate any artifacts introduced by the three - point measurements ( red in fig .
f2 ) , we also perform the same calculations of velocity and linewidth using synthetic data at considerably higher spectral resolution ( blue ) .
we find that there is no significant difference in @xmath11 or @xmath8 for both spectral resolutions except in the case of large thread numbers where the profiles can become non - gaussian . in summary , it is clear from fig .
f2 that our model can reproduce both the low amplitudes of the resolved alfvnic waves and large _ temporally invariant _ non - thermal line broadening observed with comp . extending this computation to a broader range of models
is straightforward .
the panels of fig . f3 show the variation in @xmath11 ( a ) and @xmath20 ( b ) for a range of thread numbers and input wave amplitudes where the other thread parameters remain as described above .
the white ( dot - dashed and solid ) lines show the 15 and 45km / s values of @xmath8 respectively , while the black ( dot - dashed and solid ) lines correspond to the 0.25 and 1.5km / s values of @xmath11 respectively .
these bounds encapsulate more than 75% of the analyzed pixels in panels b and c of fig .
f1 . note that iso - contours of @xmath11 are approximately quadratic functions of the ( log of the ) thread number and input wave amplitude , whereas the measured linewidths almost directly reflect the input wave amplitude of the model ( for thread numbers larger than @xmath030 ) .
one of the most stringent constraints from comp for our model is the correlation and peculiarly shaped correlation wedge that we found in fig .
f1d between the @xmath11 and @xmath8 .
we show in fig .
f4 that we can easily ( and naturally ) reproduce such a correlation of @xmath11 and @xmath8 and the wedge by using the data from the simple model presented in fig .
we use the line - enclosed regions in the panels of fig f3 ( which are driven by comp measurements ) , and apply other simple selection rules on the parameter set to best recover the correlation wedge .
we find that the shape of the wedge is sensitive to the number of threads along the line of sight , the minimum wave amplitude and any additional broadening not due to swaying motions .
panel a shows the relationship between @xmath11 and @xmath8 when the minimum wave amplitude is allowed to be zero . to match the cut - off of the wedge at 17km / s shown in the observations of fig .
f1 we need to add 17km / s of additional broadening .
the maximum and minimum number of threads determines the lower and upper slopes of the wedge .
while the slopes of the wedge are generally well reproduced , the low end of the wedge is too pointed and the scatter in @xmath11 does not fill the wedge .
this implies that this parameter set is inadequate to reproduce the observed relationship .
similarly , in panel b , we have omitted the additional broadening but , to match the observed low - end cut - off at 17km / s we must impose a minimum wave amplitude of 23km / s peak - to - peak ( which is equivalent to 16km / s rms ) .
this parameter set is also inadequate : the slope of the wedge is too shallow and it does not fill the wedge at values of high @xmath11 and @xmath8 . finally ,
in panel c , we show the optimal parameter set for this observation .
the best match fills the wedge and requires a minimum wave amplitude of 14km / s peak - to - peak ( which is equivalent to 10km / s rms ) and an additional broadening of 14km / s rms .
the high @xmath8 cut - off would indicate that the maximum wave amplitude present is 40km / s ( or 56km / s peak - to - peak ) .
the upper bound of the wedge is determined by pixels that have low thread numbers where we find the best match is at 125 threads per voxel .
the lower bound of the wedge is determined by the highest number of threads in a voxel where the best match is between 1200 and 1300 threads . in summary ,
the correlation found from comp comes about naturally if the observations are dominated by superposition of vigorous alfvnic motions ( with properties similar to sot and aia ) along a los with a large and variable number of independently oscillating structures . and
@xmath8 values used to reproduce the variance observed in panel d of fig .
the red dashed lines are taken from the bounds of the correlation wedge in fig .
panel a has no minimum wave amplitude , minimum and maximum thread numbers of 110 and 1200 ( respectively ) , but requires 17km / s of additional broadening to match the observed cut - off velocity of the wedge .
panel b has non additional broadening , minimum and maximum thread numbers of 300 and 1500 ( respectively ) , but requires 23km / s minimum wave amplitude to match the observed cut - off velocity of the wedge .
panel c is the optimal solution with 14km / s additional broadening , a 14km / s minimum wave amplitude and thread limits as shown .
[ f4],width=321 ]
to understand the low values of the resolved comp wave amplitude we have constructed a simple monte - carlo model with only two variables : the input wave amplitude , and the number of emitting threads in a voxel .
our model can reproduce the comp observations , even though it is based on the much larger wave amplitudes observed with sot and aia : comp measurements will see decreasing wave amplitudes and increasing linewidths with increasing complexity ( superposition ) of the emission along the los .
we also see that the linewidths approach temporal invariance within the accuracy of the gaussian fit - this invariance was noted by @xcite .
we note that the movies supporting fig .
f1 show that in the less dense regions ( in coronal holes ) , the linewidths do show some periodicity with time .
this would appear to be consistent with the lower panels of fig .
f2 which show low amplitude variations of linewidths where the number of threads is smaller .
the earlier deduction of complete temporal invariance @xcite may have been biased by the study of a single sub - region .
we have also reproduced the wedge relationship between @xmath11 and @xmath8 .
this relationship comes about naturally if one limits the input wave amplitudes and assumes a minimum and maximum number of independently oscillating threads along the los . under the assumptions of our simplistic model
, this would imply that observed comp wave amplitudes of @xmath01km / s are actually hiding the true wave amplitudes that range between 25 and 56km / s .
further , the measured linewidths show an almost one - to - one relationship with the amplitude of the waves present ( fig f3b ) , thus providing , in principle , an almost direct measure of the unresolved low - frequency wave energy .
we note that the exact calibration between line broadening and wave amplitude of swaying motions is actually dependent on several assumptions ( see below ) .
unresolved flows along the los likely contribute to the non - thermal line broadening observed with comp . in principle
, such flows could even cause the correlation found in fig .
f1d , as long as the flows are temporally varying ( @xmath11 ) . however , this is unlikely for a couple of reasons .
first , the @xmath11 is not dominated by flows but by waves @xcite .
second , as shown in fig .
f5 , the upward - downward directed wave power ratio of @xcite shows the same correlation with the @xmath8 as @xmath11 . dividing the coronal cavity into its southern ( foreground ) and northern ( background ) portions by the dashed radial line we plot the correlation of @xmath8 and the directional wave power ratio . in both regions
the relationship is linear , which supports the scenario in which @xmath8 is dominated by the propagating wave motions of the plasma and not by bulk coronal flows along the los .
our simplistic model has several limitations .
for example , we assume that the transverse motions are dominated by swaying , and not torsional , motions . the latter would not cause resolved doppler shift variations with time .
it is possible that the additional broadening of 14km / s ( uncorrelated with @xmath11 ) required to reproduce the correlation wedge is a signature of such torsional motions
. this would be consistent with recent observations of peak - to - peak amplitudes in spicules of order 20 - 30 km / s @xcite .
it is also possible , as recent studies suggest , that some of the swaying motions couple to torsional modes .
if such a coupling mechanism occurs in the corona ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ) , the amplitude of the torsional modes could be correlated with that of the swaying motions , so that the line broadening caused by torsional modes could actually be part of the broadening that _ is _ correlated with @xmath11 . in this case , torsional motions would _ not _ cause the additional broadening ( which instead could be caused by unresolved flows along the los ) .
if this were the case , the observed correlation between @xmath11 ( from swaying ) and @xmath20would not necessarily all be caused by los superposition . in this scenario , the relative contributions of swaying and torsional motions likely change with height ( with increasing contributions from torsional motions higher up ) , because the coupling mechanism suggested by @xcite occurs over a finite time /distance during which waves propagate upward from the loop footpoints .
taken at face value , our model would predict a very large number of independently oscillating threads along the los ( @xmath01 solar radius ) contained in one comp pixel ( 4,500 km ) .
it may well be the case that the corona is extremely finely structured .
however , the number of threads predicted by our model may be very different from the number of threads in the corona .
this is because the exact number will critically depend on the mix of torsional and swaying motions , and on the ( perpendicular ) coherence of the waves .
the latter is unknown : if phase mixing plays a significant role , wave amplitudes seen by comp can , in principle , be quenched much more efficiently than the completely independently oscillating threads assumed here .
our data seem to exclude a scenario whereby high frequency alfvnic waves ( invoked by several models for the solar wind ) dominate non - thermal line broadening ( at least for heights @xmath250 mm ) .
if the wave generation process low down predominantly created high frequency waves , with a small amount of wave power at low frequencies , one could in principle recover the correlation wedge .
however , the aia and sot observations of strong , low frequency alfvnic swaying at low heights argue against such a scenario .
ncar is sponsored by the national science foundation .
comp data can be found at http://mlso.hao.ucar.edu/. _
hinode _ is a japanese mission developed and launched by isas / jaxa , with naoj as a domestic partner and nasa and stfc ( uk ) as international partners .
we acknowledge support from nasa contracts nnx08ba99 g , nnx11an98 g , nnm12ab40p , nng09fa40c ( _ iris _ ) , and nnm07aa01c ( _ hinode _ ) .
alfvn , h. , 1947 , mnras , 107 , 211 akiyama , s. , doschek , g. a. & mariska , j. t. , 2005 , , 623 , 540 banerjee , d. , et al .
, 1998 , , 339 , 208 belcher , j. w. , 1971 , , 168 , 505 chae , j. , schhle , u. , lemaire , p. , , 505 , 957 cranmer , s. r. & van ballegooijen , a. a. , 2005 , , 156 , 265 cranmer , s. r. , van ballegooijen , a. a. , & edgar , r. j. , 2007 , , 171 , 520 de moortel , i. , & pascoe , d. j. 2012 , , 746 , 31 de pontieu , b. , et al . , 2007a , science , 318 , 1574 , b. , et al . ,
2007b , pub .
jap . , 59 , 65 de pontieu , b. , mcintosh , s. w. , carlsson , m. , et al . 2011 , science , 331 , 55 de pontieu , b. , carlsson , m. , rouppe van der voort , l. h. m. , et al .
2012 , , 752 , l12 hassler , d. m. , et al . , 1990 , , 348 , l77 hassler , d. m. , moran , t.g . , 1994 , sp .
, 70 , 373 hansteen , v. h. , leer , e. , 1995 . , , 100 , 21577 hollweg , j.v . , 1978 , , 56 , 305 hollweg , j.v .
, 1981 , , 70 , 25 ionson , j.a . , 1978 , , 226 , 650 kosugi , t. , et al . 2007 , , 243 , 3 lemen , j. r. , title , a. m. , akin , d. j. , et al . 2011 , , 172 mcintosh , s. w. , de pontieu , b. & tarbell , t.d . , 2008 , , 673 , l219 mcintosh , s. w. , de pontieu , b. , carlsson , m. , et al .
2011 , , 475 , 477 melrose , d. b. 1977 , aust .
j. of phys .
, 30 , 495 moran , t. , g. , 2003 , , 374 , l9 moran , t. , g. , 2003 , , 598 , 657 pascoe , d. j. , wright , a. n. , & de moortel , i. 2011 , , 731 , 73 peter , h. , 1999a , , 516 , 490 peter , h. , 1999b , , 522 , l77 peter , h. , 2000a , , 360 , 761 peter , h. & vocks , c. , 2003 , , 411 , l481 ruderman , m.s .
, & roberts , b. , 2002 , , 577 , 475 singh , j. , sakurai , t. & ichimoto , k. , 2006 , , 639 , 475 singh , j. , sakurai , t. , ichimoto , k. , muneer , s. & raveendran , a. v. , 2006 , , 236 , 245 tomczyk , s. , et al . , 2007 , science , 317 , 1192 tomczyk , s. , et al . , 2008 , , 247 , 411 tomczyk , s. , & mcintosh , s. w. 2009 , , 697 , 1384 tsuneta , s. , et al . , 2008 , , 249 , 167 verdini , a. & velli , m. , 2007 , , 662 , 701 van ballegooijen , a. a. , asgari - targhi , m. , cranmer , s. r. , & deluca , e. e. 2011 , , 736 , 3 verth , g. , terradas , j. , and goossens , m. , 2010 , 718 , l102 withbroe , g. l. & noyes , r. , w. , 1977 , ann .
astrophys . , 15 , 363 | the discovery of ubiquitous low - frequency ( 3 - 5mhz ) alfvnic waves in the solar chromosphere ( with hinode / sot ) , and corona ( with comp and the _ solar dynamics observatory _ , _ sdo _ ) has provided some insight into the non - thermal energy content of the outer solar atmosphere .
however , many questions remain about the true magnitude of the energy flux carried by these waves . here
we explore the apparent discrepancy in the resolved coronal alfvnic wave amplitude ( @xmath00.5km / s ) measured by the coronal multi - channel polarimeter ( comp ) compared to those of the _ hinode _ and the _ solar dynamics observatory _
( _ sdo _ ) near the limb ( @xmath020km / s ) .
we use a blend of observational data and a simple forward model of alfvnic wave propagation to resolve this discrepancy and determine the alfvnic wave energy content of the corona .
our results indicate that enormous line - of - sight superposition within the coarse spatio - temporal sampling of comp hides the strong wave flux observed by _
hinode _ and _ sdo _ and leads to the large non - thermal line broadening observed .
while this scenario has been assumed in the past , our observations with comp of a strong correlation between the non - thermal line broadening with the low amplitude , low frequency alfvnic waves observed in the corona provide the first direct evidence of a wave - related non - thermal line broadening . by reconciling the diverse measurements of alfvnic waves we establish large coronal non - thermal linewidths as direct signatures of the hidden , or `` dark '' , energy content in the corona , and provide preliminary constraints on the energy content of the wave motions observed . |
the domain - wall fermion formulation of qcd ( dwqcd ) @xcite is expected to realize exact chiral symmetry on the lattice without species doubling at finite lattice spacing .
this represents an appealing possibility , particularly for investigations of problems such as weak matrix elements sensitive to chiral symmetry @xcite .
therefore , a number of studies have been made @xcite .
simulations in dwqcd , however , requires considerable computing power .
even if the size of the extra dimension @xmath4 can be taken as small , i.e. @xmath5 , lattice spacings much finer than @xmath6 gev will be difficult to simulate even in quenched qcd .
hence simulations on coarse lattices down to @xmath7 gev will be needed for reliable continuum extrapolations .
a first step in dwqcd at such a strong coupling will be to ensure the existence of chiral zero modes . here
we report our preliminary results on this problem .
our study is made for the plaquette action , and also for an rg - improved action @xcite to examine the effects of reduced discretization errors .
we employ the fermion action identical to that in ref .
@xcite , with the domain wall height @xmath8 and bare quark mass @xmath9 .
the gauge coupling is chosen to be @xmath10 for the plaquette action and @xmath11 for the rg action ; these values correspond to the scale @xmath12 gev determined from the string tension .
simulations are made on an @xmath1 lattice with @xmath13 and 50 . in the free fermion case the chiral zero mode exists over the range @xmath14 . since this range will be shifted in the interacting theory , we employ @xmath15 and 2.5 . for each of these values of @xmath8 , we take @xmath16 and 0.03 and calculate the pion mass for both degenerate and non - degenerate quark and antiquark pairs .
for each parameter point we have typically @xmath17 configurations .
= 7.6 cm = 7.6 cm in fig .
[ fig : pi2-mfns ] the pion mass squared @xmath18 is plotted as a function of the averaged bare quark mass @xmath19 at @xmath20 in the case of the plaquette gauge action .
since the linearity of @xmath18 in @xmath19 is well satisfied , we adopt a linear chiral extrapolation in @xmath19 . as it can be seen in the fig .
[ fig : pi2-mfns ] , we find a non - zero value for @xmath18 at @xmath21 for each @xmath4 .
this non - zero value , however , is expected to vanish exponentially as @xmath22 . to see whether this occurs , we plot @xmath18 at @xmath23 by thick square symbols in fig .
[ fig : pi2-ns](a ) as a function of @xmath4 .
fitting these points by the form @xmath24 yields the dotted line with @xmath25 .
an alternative fit , allowing a constant , @xmath26 gives the solid line with @xmath27 and @xmath28 .
clearly the latter fit better reproduces the behavior of our data .
a similar phenomenon has been previously reported also at @xmath29 @xcite . in order to confirm the existence of a non - zero @xmath30
, we attempt to interchange the order of the limits @xmath31 and @xmath32 . as shown by solid lines going through open circles in fig .
[ fig : pi2-ns](a ) , we first make a fit of form @xmath33 to make an @xmath32 extrapolation for each value of @xmath19 .
the results , shown by solid circles in fig . [ fig : pi2-mfns ] , exhibits a linear behavior in @xmath19 . a linear chiral extrapolation
@xmath34 then yields @xmath35 , which agrees well with the value of @xmath30 previously obtained .
the commutativity of the two limits is summarized in fig .
[ fig : pi2-ns](a ) , where the two values obtained for @xmath36 are marked by `` x '' and `` y '' .
the above analyses strongly support the conclusion that @xmath18 at @xmath21 does not vanish in the limit @xmath37 at @xmath38 1 gev for the 4-dimensional lattice size of @xmath39 .
we find this conclusion to apply as well to the rg - improved gauge action , as shown in fig .
[ fig : pi2-ns](b ) . in fig .
[ fig : pi2-m ] the residual @xmath40 is plotted as a function of @xmath8 for the plaquette and rg - improved gauge actions .
we observe that a non - vanishing residual remains under the variation of @xmath8 for both actions .
an improvement may be seen for the rg action , however , in that the magnitude of the residual is reduced by about a factor of two compared to that for the plaquette action .
finally the decay rate @xmath41 extracted from the fits is shown in fig.[fig : dlate - mf ] as a function of @xmath19 at @xmath20 .
for both actions , the chiral extrapolation of @xmath41 from @xmath42 ( open circles ) smoothly agrees with the one directly obtained at @xmath23 assuming @xmath43 ( filled circle ) , but disagrees with the value from the @xmath44 fit .
furthermore the values of @xmath41 varies little with @xmath19 .
this is consistent with the expectation that the decay rate is governed by the transfer matrix in the extra dimension , which is independent of @xmath9 .
our study of domain - wall qcd with the plaquette gauge action has shown that the pion mass in the chiral limit remains non - zero for an infinitely sized extra dimension at a strong gauge coupling corresponding to @xmath0 1 gev on a @xmath39 lattice .
we have also found that this conclusion remains unchanged for the rg - improved gauge action .
one possibility for the origin of a non - zero pion mass is finite spatial size effects , which is being checked by increasing the spatial size from 12 to 16 .
another possibility is that it is an artifact of the linear chiral extrapolation , which does not take into account the possible presence of quenched chiral logarithms . to explore this possibility
, the chiral breaking term in the axial ward - takahashi identity @xcite , which we expect to be free from the quenched singularities , is currently being investigated .
finally it is possible that no chiral zero modes exist at @xmath0 1 gev , and therefore the domain - wall formalism fails to realize chiral symmetry in the region of strong coupling . a plausible explanation for this failure might be that the range of @xmath8 for zero modes to exist disappears for stronger couplings , where there is no gap of eigenvalues for the 4-dimensional wilson operator @xcite .
this work is supported in part by the grants - in - aid of ministry of education ( nos .
09304029 , 10640246 , 10640248 , 11640250 , 10740107 , 11640294 , 11740162 ) .
se , ti , kn and yt are jsps research fellows .
aak and tm are supported by the research for the future program of jsps .
9 d. kaplan , phys .
b288 ( 1992 ) 342 . v. furman and y. shamir , nucl .
b439 ( 1995 ) 54 .
t. blum and a. soni , phys .
d56 ( 1997 ) 174 ; phys . rev
. lett .
79 ( 1997 ) 3595 .
t. blum , nucl .
b ( proc . suppl . ) 73 ( 1999 ) 167 .
univ . of tsukuba
report uthep-118 ( 1983 ) , unpublished .
p. chen _ et al .
_ , nucl .
b ( proc . suppl . ) 73 ( 1999 ) 204 .
r. setoodeh , c. t. h. davies and i. m. barbour , phys .
b213 ( 1988 ) 195 . | we investigate the existence of chiral zero modes at @xmath0 1 gev in quenched domain - wall qcd .
simulations are carried out for the plaquette and an rg - improved gauge actions on a @xmath1 lattice with @xmath2 .
we find that the pion mass in the chiral limit remains non - vanishing as @xmath3 for both gauge actions .
possible origins of this non - vanishing pion mass are discussed . |
generalized matrix functions were first introduced by hawkins and ben - israel in @xcite in order to extend the notion of a matrix function to rectangular matrices . essentially , the definition is based on replacing the spectral decomposition of @xmath0 ( or the jordan canonical form , if @xmath0 is not diagonalizable ) with the singular value decomposition , and evaluating the function at the singular values of @xmath0 , if defined . while it is likely that this definition was inspired by the analogy between the inverse and the moore
penrose generalized inverse , which is well established in numerical linear algebra , @xcite is a purely theoretical paper and does not mention any potential applications or computational aspects .
the paper appears to have gone largely unnoticed , despite increasing interest in matrix functions in the numerical linear algebra community over the past several years ; for instance , it is not cited in the important monograph by higham @xcite .
while it is likely that the perceived scarcity of applications is to blame ( at least in part ) for this lack of attention , it turns out that generalized matrix functions do have interesting applications and have actually occurred in the literature without being recognized as such ; see section [ sec : app ] for some examples . in this paper
we revisit the topic of generalized matrix functions , with an emphasis on numerical aspects . after reviewing the necessary background and definitions
, we consider a few situations naturally leading to generalized matrix functions . moving on to numerical considerations
, we develop several computational approaches based on variants of golub kahan bidiagonalization to compute or estimate bilinear forms involving generalized matrix functions , including entries of the generalized matrix function itself and the action of a generalized matrix function on a vector .
we further consider block variants of golub kahan bidiagonalization which can be used to evaluate matrix - valued expressions involving generalized matrix functions .
numerical experiments are used to illustrate the performance of the proposed techniques on problems arising in the analysis of directed networks .
in this section we review a few basic concepts from linear algebra that will be used throughout the paper , mostly to set our notation , and recall the notion of generalized matrix function .
let @xmath1 and let @xmath2 be the rank of @xmath0 .
we can factor the matrix @xmath0 as @xmath3 using a singular value decomposition ( svd ) .
the matrix @xmath4 is diagonal and its entries @xmath5 are ordered as @xmath6 , where @xmath7 .
the @xmath2 positive @xmath8 are the singular values of @xmath0 .
the matrices @xmath9\in{{\mathbb{c}}}^{m\times m}$ ] and @xmath10\in{{\mathbb{c}}}^{n\times n}$ ] are unitary and contain the left and right singular vectors of @xmath0 , respectively .
it is well known that the matrix @xmath11 is uniquely determined , while @xmath12 and @xmath13 are not .
if @xmath0 is real , then @xmath12 and @xmath13 can be chosen to be real . from the singular value decomposition of a matrix @xmath0
it follows that @xmath14 and @xmath15 .
thus , the singular values of a matrix @xmath0 are the square roots of the positive eigenvalues of the matrix @xmath16 or @xmath17 . moreover , the left singular vectors of @xmath0 are the eigenvectors of the matrix @xmath16 , while the right singular vectors are the eigenvectors of the matrix @xmath17 .
the singular values of a matrix also arise ( together with their opposites ) as the eigenvalues of the hermitian matrix @xmath18 this can be easily seen for the case @xmath19 ; indeed , under this hypothesis , the spectral factorization of @xmath20 is given by @xcite : @xmath21 consider now the matrices @xmath22 and @xmath23 which contain the first @xmath2 columns of the matrices @xmath12 and @xmath13 , respectively , and let @xmath24 be the @xmath25 leading block of @xmath11 .
then a _ compact svd _
( _ csvd _ ) of the matrix @xmath0 is @xmath26 there are several equivalent ways to define @xmath27 when @xmath28 is a square matrix .
we recall here the definition based on the jordan canonical form . for a comprehensive study of matrix functions ,
we refer to @xcite .
let @xmath29 be the set of distinct eigenvalues of @xmath0 and let @xmath30 denote the _ index _ of the @xmath31th eigenvalue , i.e. , the size of the largest jordan block associated with @xmath32 . recall that a function @xmath33 is said to be _ defined on the spectrum of @xmath0 _ if the values @xmath34 exist for all @xmath35 and for all @xmath36 , where @xmath37 is the @xmath38th derivative of the function and @xmath39 .
* definition 1.2)[def : fun ] let @xmath33 be defined on the spectrum of @xmath40 and let @xmath41 be the jordan canonical form of the matrix , where @xmath42 @xmath43 , and @xmath44 is nonsingular .
then @xmath45 where @xmath46 if the matrix @xmath0 is diagonalizable , then the jordan canonical form reduces to the spectral decomposition : @xmath47 with @xmath48 . in such case ,
@xmath49 .
if the function @xmath33 has a taylor series expansion , we can use it to describe the associated matrix function , provided that the eigenvalues of the matrix @xmath0 satisfy certain requirements .
* theorem 4.7 ) let @xmath28 and suppose that @xmath33 can be expressed as @xmath50 with radius of convergence @xmath51 . then @xmath27 is defined and is given by @xmath52 if and only if each of the distinct eigenvalues of @xmath0 @xmath29 satisfies one of the following : * @xmath53 ; * @xmath54 and the series for @xmath55 is convergent at the point @xmath56 , @xmath57 . in @xcite the authors considered the problem of defining functions of rectangular matrices .
their definition relies on the following generalization of the svd .
[ thm : ci ] let @xmath58 be a matrix of rank @xmath2 and let @xmath59 be any complex numbers satisfying @xmath60 where @xmath61 are the positive eigenvalues of @xmath16 . then there exist two unitary matrices @xmath62 and @xmath63 such that @xmath64 has entries : @xmath65 from this theorem it follows that , once the non - zero entries of @xmath66 are fixed , @xmath0 can be written as @xmath67 where @xmath68 is the leading @xmath25 block of @xmath69 and the matrices @xmath70 and @xmath71 consist of the first @xmath2 columns of the matrices @xmath72 and @xmath73 , respectively .
in this paper we do not make use of the extra degrees of freedom provided from this ( slightly ) more general svd of @xmath0 , and we assume that @xmath74 for all @xmath75 .
this assumption ensures that the decompositions in coincide with the svd and csvd of the matrix @xmath0 , respectively .
in particular , @xmath76 , @xmath77 , and @xmath78 . all the definitions and results presented in the remaining of this section and in the next one can be extended to the case when the coefficients @xmath79 do not necessarily coincide with the singular values , but satisfy the hypothesis of theorem [ thm : ci ] .
[ def : gf ] let @xmath1 be a rank @xmath2 matrix and let @xmath80 be its csvd .
let @xmath81 be a scalar function such that @xmath82 is defined for all @xmath75 .
the generalized matrix function @xmath83 induced by @xmath33 is defined as @xmath84 where @xmath85 is defined for the square matrix @xmath86 according to definition [ def : fun ] as @xmath87 as already mentioned in the introduction , generalized matrix functions arise , for instance , when computing @xmath88 , where @xmath20 is the matrix defined in .
indeed , if one uses the description of matrix function in terms of power series @xmath89 , it is easy to check that , within the radius of convergence : @xmath90 where @xmath91 if @xmath92 and the matrix @xmath28 is hermitian positive semidefinite , then the generalized matrix function @xmath93 reduces to the standard matrix function @xmath27 .
if the more general decomposition of theorem [ thm : ci ] is used instead , then the generalized matrix function reduces to @xmath27 if @xmath92 and the matrix @xmath28 is normal , as long as @xmath33 is defined on the @xmath94 ; see @xcite . the moore
penrose pseudo - inverse of a matrix @xmath1 , denoted by @xmath95 , can be expressed as @xmath96 , where @xmath97 .
equivalently , @xmath98 when @xmath97 .
hence , there is a slight disconnect between the definition of generalized matrix function and that of generalized inverse .
this small issue could be addressed by defining a generalized matrix function corresponding to the scalar function @xmath33 as @xmath99 , so that the generalized matrix function of an @xmath100 matrix is an @xmath101 matrix , and @xmath102 when @xmath97 . however , doing so would lead to the undesirable property that @xmath103 for @xmath104 , as well as other problems .
in this section we review some properties of generalized matrix functions and we summarize a few new results . letting @xmath105 and @xmath106
, we can write @xmath107 and thus it follows that @xmath108 ( sums and products of functions @xcite ) . let @xmath109 be scalar functions and let @xmath110 be the corresponding generalized matrix functions . then : * if @xmath111 , then @xmath112 ; * if @xmath113 , then @xmath114 ; * if @xmath115 , then @xmath116 ; * if @xmath117 , then @xmath118 . in the following
we prove a few properties of generalized matrix functions .
[ thm : prop ] let @xmath1 be a matrix of rank @xmath2 .
let @xmath81 be a scalar function and let @xmath119 be the induced generalized matrix function , assumed to be defined at @xmath0 .
then the following properties hold true . * @xmath120^ * = { f^{\diamond}}(a^*)$ ] ; * let @xmath121 and @xmath122 be two unitary matrices , then @xmath123y$ ] ; * if @xmath124 , then @xmath125 * @xmath126 , where @xmath127 is the @xmath128 identity matrix and @xmath129 is the kronecker product ; * @xmath130 . * from it
follows that @xmath131 , and thus @xmath132^ * = [ { f^{\diamond}}(a)]^*.\ ] ] * the result follows from the fact that unitary matrices form a group under multiplication and that the rank of a matrix does not change under left or right multiplication by a nonsingular matrix @xcite .
indeed , the matrix @xmath133 has rank @xmath2 and thus @xmath134^ * \\ & = x u_r f(\sigma_r ) v^*_r y = x{f^{\diamond}}(a)y .
\end{aligned}\ ] ] where @xmath135 and @xmath136 are the matrices containing the first @xmath2 columns of @xmath137 and @xmath138 , respectively .
* let @xmath139 be the csvd of the rank-@xmath140 matrix @xmath141 for @xmath142 .
then @xmath143 , where @xmath144 is a diagonal matrix whose diagonal entries are ordered ( via the permutation matrix @xmath145 ) in non - increasing order , and @xmath146 @xmath147 from the definition of generalized matrix function and from some basic properties of standard matrix functions @xcite it follows that @xmath148 * the result follows from ( iii ) and the fact that @xmath149 is a @xmath150 diagonal block matrix with @xmath151 copies of @xmath0 on the main diagonal .
* it follows from ( iv ) and from the fact that for two general matrices @xmath1 and @xmath152 , there exist two permutation matrices @xmath153 and @xmath154 called _ commutation matrices _ such that @xmath155 ( see ( * ? ? ?
3 ) ) . the following theorem provides a result for the composition of two functions .
( composite functions ) let @xmath1 be a rank-@xmath2 matrix and let @xmath156 be its singular values .
assume that @xmath157 and @xmath158 are two scalar functions such that @xmath159 and @xmath160 exist for all @xmath75 .
let @xmath161 and @xmath162 be the induced generalized matrix functions .
moreover , let @xmath81 be the composite function @xmath163 .
then the induced matrix function @xmath119 satisfies @xmath164 let @xmath165 . since @xmath159 for all @xmath75 , this matrix has rank @xmath2 .
we want to construct a csvd of the matrix @xmath166 .
let thus @xmath167 be a permutation matrix such that the matrix @xmath168 has diagonal entries ordered in non - increasing order .
then it follows that a csvd of the matrix @xmath166 is given by @xmath169 , where @xmath170 and @xmath171 have orthonormal columns .
it thus follows that @xmath172 the following result describes the relationship between standard matrix functions and generalized matrix functions .
[ thm : genfun ] let @xmath1 be a rank-@xmath2 matrix and let @xmath81 be a scalar function .
let @xmath119 be the induced generalized matrix function .
then @xmath173 or , equivalently , @xmath174 the two identities are an easy consequence of the fact that @xmath175 and @xmath176 for @xmath75 .
let @xmath1 be a rank-@xmath2 matrix and let @xmath81 and @xmath158 be two scalar functions such that @xmath93 and @xmath177 are defined .
then @xmath178 from @xmath179 it follows @xmath180 and @xmath181 ; thus @xmath182
as mentioned in the introduction , generalized matrix functions ( in the sense of hawkins and ben - israel ) have appeared in the literature without being recognized as such .
here we discuss a few examples that we are aware of .
no doubt there have been other such instances . in @xcite ,
the authors address the problem of computing functions of real skew - symmetric matrices , in particular the evaluation of the product @xmath183 for a given skew - symmetric matrix @xmath0 and vector @xmath184 using the lanczos algorithm .
the authors observe that any @xmath185 with @xmath186 is orthogonally similar to a matrix of the form @xmath187 where @xmath166 is lower bidiagonal of order @xmath188 . as a consequence ,
if @xmath189 is an svd of @xmath166 , the matrix exponential @xmath190 is orthogonally similar to the matrix @xmath191 where the matrix in the upper right block is precisely @xmath192 .
the authors of @xcite develop computational techniques for the matrix exponential based on ( [ eq : lopez ] ) .
we also mention that in the same paper the authors derive a similar expression , also found in @xcite , for the exponential of the symmetric matrix @xmath20 given in ( [ eq : cala ] ) .
these expressions are extended to more general matrix functions in @xcite , where they are used to investigate the off - diagonal decay of analytic functions of large , sparse , skew - symmetric matrices .
furthermore , in @xcite it is shown how these ideas can be used to develop efficient geometrical integrators for the numerical solution of certain hamiltonian differential systems . in @xcite ,
the authors consider the problem of detecting ( approximate ) directed bipartite communities in directed graphs .
consideration of alternating walks in the underlying graph leads them to introducing a non - standard matrix function " of the form @xmath193 where @xmath0 is the adjacency matrix of the graph . using @xmath194 this expression
is readily recognized to be equivalent to @xmath195 which is a mixture " of the standard matrix function @xmath196 and the generalized matrix function @xmath197 .
as mentioned , generalized hyperbolic matrix functions were also considered in @xcite in the context of directed networks , also based on the notion of alternating walks in directed graphs . in @xcite , the action of generalized matrix functions on a vector of all ones was used to define certain centrality measures for nodes in directed graphs ; here the connection with the work of hawkins and ben - israel was explicitly made .
finally , we mention that generalized matrix functions arise when _ filter factors _
are used to regularize discrete ill - posed problems ; see , e.g. , @xcite .
the computation of the generalized matrix functions defined as in definition [ def : gf ] requires the knowledge of the singular value decomposition of @xmath0 . when @xmath198 and @xmath188 are large , computing the svd may be unfeasible .
moreover , in most applications it is not required to compute the whole matrix @xmath93 ; rather , the goal is often to estimate quantities of the form @xmath199 or to compute the action of the generalized matrix function on a set of @xmath151 vectors , i.e. , to evaluate @xmath200 , usually with @xmath201 .
for example , computing selected columns of @xmath93 reduces to the evaluation of @xmath200 where @xmath202 consists of the corresponding columns of the identity matrix @xmath203 , and computing selected entries of @xmath93 requires evaluating @xmath204 where @xmath44 contains selected columns of the identity matrix @xmath205 .
the problem of estimating or giving bounds on such quantities can be tackled , following @xcite , by using gauss - type quadrature rules . as usual in the literature
, we will first analyze the case @xmath206 ; the case of @xmath207 will be dealt with in section [ sec : approach_bl ] .
it is known that in certain cases gauss - type quadrature rules can be used to obtain lower and upper bounds on bilinear forms like @xmath209 , where @xmath27 is a ( standard ) matrix function and @xmath210 .
this is the case when @xmath33 enjoys certain monotonicity properties .
recall that a real - valued function @xmath33 is _ completely monotonic _
_ ) on an interval @xmath211 $ ] if it is continuous on @xmath212 and infinitely differentiable on @xmath213 with @xmath214 where @xmath215 denotes the @xmath188th derivative of @xmath33 and @xmath39 .
if @xmath33 is completely monotonic on an interval containing the spectrum of @xmath210 , then one can obtain lower and upper bounds on quadratic forms of the type @xmath216 and from these lower and upper bounds on bilinear forms like @xmath217 with @xmath218 . for a general @xmath33 ,
on the other hand , gaussian quadrature can only provide estimates of these quantities .
similarly , in order to obtain bounds ( rather than mere estimates ) for bilinear expressions involving generalized matrix functions , we need the scalar functions involved in the computations to be completely monotonic .
we will be applying our functions to diagonal matrices that contain the singular values of the matrix of interest .
thus , in our framework , the interval on which we want to study the complete monotonicity of the functions is @xmath219 .
we briefly recall here a few properties of c. m. functions ; see , e.g. , @xcite and references therein for systematic treatments of complete monotonicity .
[ lemma3 ] if @xmath220 and @xmath221 are completely monotonic functions on @xmath212 , then * @xmath222 with @xmath223 is completely monotonic on @xmath212 ; * @xmath224 is completely monotonic on @xmath212 .
* theorem 2)[lemma1 ] let @xmath220 be completely monotonic and let @xmath221 be a nonnegative function such that @xmath225 is completely monotonic .
then @xmath226 is completely monotonic . using these lemmas
, we can prove the following useful result .
if @xmath33 is completely monotonic on @xmath227 , then @xmath228 is completely monotonic on @xmath227 .
let @xmath229 ; then by lemma [ lemma3 ] ( ii ) we know that @xmath230 is completely monotonic on @xmath219 if both @xmath231 and @xmath232 are completely monotonic on @xmath212 .
the function @xmath233 is positive on the interval @xmath227 ; moreover , it is such that its first derivative @xmath234 is completely monotonic on @xmath212 .
therefore , from lemma [ lemma1 ] it follows that if @xmath33 is c.m . , then @xmath231 is .
similarly , since @xmath229 is completely monotonic , @xmath232 is completely monotonic .
this concludes the proof . in the following ,
we propose three different approaches to approximate the bilinear forms of interest .
the first approach exploits the results of theorem [ thm : genfun ] to describe @xmath235 as a bilinear form that involves standard matrix functions of a tridiagonal matrix .
the second approach works directly with the generalized matrix function and the moore
penrose pseudo - inverse of a bidiagonal matrix .
the third approach first approximates the action of a generalized matrix function on a vector and then derives the approximation for the bilinear form of interest . when the function @xmath236 that defines @xmath237 is c.m .
, then gauss - type quadrature rules can be used to derive upper and lower bounds for the quantities of interest .
it is straightforward to see by using that a bilinear form involving a generalized matrix function can be written as @xmath238 where @xmath239 , and @xmath240 . using the equalities in one
can see that these quantities can also be expressed as bilinear forms involving functions of the matrices @xmath241 and @xmath242 , respectively .
more in detail , one obtains @xmath243 @xmath244 where in both cases @xmath245 . in the following we focus on the case described by .
the discussion for the case described by follows the same lines .
note that if @xmath246 are vectors such that @xmath247 , then we can use the _ polarization identity _
@xcite : @xmath248 \ ] ] to reduce the evaluation of the bilinear form of interest to the evaluation of two symmetric bilinear forms . for this reason
, the theoretical description of the procedure to follow will be carried out only for the case @xmath249 .
let @xmath249 be a unit vector ( i.e. , @xmath250 ) .
we can rewrite the quantity as a riemann stieltjes integral by substituting the spectral factorization of @xmath242 : @xmath251 where @xmath252 is a piecewise constant step function with jumps at the positive eigenvalues @xmath253 of @xmath242 defined as follows : @xmath254 we use partial golub
kahan bidiagonalization @xcite of the matrix @xmath0 to find upper and lower bounds for the bilinear form described in .
after @xmath255 steps , the golub
kahan bidiagonalization of the matrix @xmath0 with initial vector @xmath256 yields the decompositions @xmath257 where the matrices @xmath258\in{{\mathbb{r}}}^{n\times\ell}$ ] and @xmath259\in{{\mathbb{r}}}^{m\times\ell}$ ] have orthonormal columns , the matrix @xmath260 is upper bidiagonal , and the first column of @xmath261 is @xmath256 .
all the @xmath262 and @xmath263 can be assumed to be nonzero @xcite . with this assumption ,
the csvd of the bidiagonal matrix @xmath264 coincides with its svd : @xmath265 where @xmath266\in{{\mathbb{r}}}^{\ell\times\ell}$ ] and @xmath267\in{{\mathbb{r}}}^{\ell\times\ell}$ ] are orthogonal , and @xmath268 .
combining the equations in leads to @xmath269 where @xmath270 denotes the lanczos vector computed at iteration @xmath271 .
the matrix @xmath272 is thus symmetric and tridiagonal and coincides ( in exact arithmetic ) with the matrix obtained when the lanczos algorithm is applied to @xmath242 . the quadratic form in
can then be approximated by using an @xmath255-point gauss quadrature rule @xcite : @xmath273 if the function @xmath274 is c.m .
, then the gauss rule provides a lower bound for , which can be shown to be strictly increasing with @xmath255 .
if the recursion formulas for the golub kahan bidiagonalization break down , that is , if @xmath275 at step @xmath255 , then the gauss quadrature rule gives the exact value ( see @xcite ) .
the following result can be easily derived from equation .
[ prop : nodes ] let @xmath276 and let @xmath277 be the bidiagonal matrix computed after @xmath255 steps of the golub
kahan bidiagonalization algorithm .
let @xmath278 for @xmath279 be the singular triplets of @xmath280 .
then the nodes of the @xmath255-point gauss quadrature rule @xmath281 are the singular values @xmath282 .
furthermore , if @xmath283 , the weights of @xmath281 are @xmath284 for @xmath279 . similarly ,
if @xmath249 , then the weights of the rule are given by @xmath285 . to provide an upper bound for when @xmath33 is c. m. , one can use a @xmath286-point gauss
radau quadrature rule with a fixed node @xmath287 ; this can be expressed in terms of the entries of the symmetric tridiagonal matrix @xmath288 as @xmath289 , where @xmath290 .
the entries of this matrix , except for the last diagonal entry , are those of @xmath291 . to compute the last diagonal entry so that @xmath292 has @xmath287 among its eigenvalues
, we proceeds as follows @xcite .
first , we compute @xmath293 ; then we set @xmath294 , where @xmath295 is the solution of the tridiagonal linear system @xmath296 .
the arithmetic mean between the @xmath255-point gauss rule @xmath281 and the @xmath286-point gauss
radau rule @xmath297 is then used as an approximation of the quadratic form @xmath298 . in this section
we provide a second approach to the approximation of bilinear forms expressed in terms of generalized matrix functions .
the following result shows how to compute the @xmath255-point gauss quadrature rule in terms of the generalized matrix function of the bidiagonal matrix @xmath264 .
two expressions are derived , depending on the starting ( unit ) vector given as input to the golub
kahan algorithm .
recall that , unless @xmath299 or @xmath300 , one has to use the polarization identity to estimate the bilinear forms of interest .
let be @xmath276 and let @xmath277 be the bidiagonal matrix computed at step @xmath255 of the golub
kahan bidiagonalization algorithm .
then , the @xmath255-point gauss quadrature rule @xmath281 is given by @xmath301 or @xmath302 let @xmath303 be a singular value decomposition of the matrix @xmath264 obtained after @xmath255 steps of the golub
kahan bidiagonalization algorithm with starting vector @xmath249 .
then , from proposition [ prop : nodes ] , it follows that @xmath304 the proof of the case when @xmath305 goes along the same lines and it is thus omitted .
the @xmath306-point gauss - radau quadrature rule @xmath307 with a fixed node @xmath308 can be expressed in terms of the entries of the bidiagonal matrix @xmath309 as @xmath310 if @xmath311 or as @xmath312 when @xmath313 .
the entries of @xmath314 , except for the last diagonal entry , are those of @xmath315 . to compute the last diagonal entry
, one has to ensure that @xmath316 is an eigenvalue of @xmath317 .
it can be easily shown that @xmath318 where @xmath319 is the solution of the tridiagonal linear system @xmath320 .
assume that we have used @xmath321 steps of the golub
kahan bidiagonalization algorithm with starting vector @xmath256 ( normalized so as to have unit norm ) to derive the matrices @xmath322 , @xmath323 , and @xmath324 such that @xmath325 .
the csvd of the bidiagonal matrix is @xmath326 , where @xmath86 is the same diagonal matrix appearing in the csvd of @xmath0 .
since @xmath322 and @xmath324 have full column rank , we know that @xmath327 , and thus we can write @xmath328 where @xmath329 and @xmath330 .
assume now that @xmath331 .
we can then truncate the bidiagonalization process and approximate @xmath332 as @xmath333 and then obtain the approximation to the bilinear form of interest as @xmath334 the quality of the approximation will depend in general on the distribution of the singular values of @xmath0 and on the particular choice of @xmath33 . generally speaking ,
if @xmath82 is much larger on the first few singular values of @xmath0 than for the remaining ones , then a small number of steps result in approximations with small relative errors .
in this section we describe two ways to compute approximations of quantities of the form , when @xmath207 .
it is known that for this kind of problem , block algorithms are generally more efficient than the separate computation of each individual entry ( or column ) of @xmath335 .
when dealing with blocks @xmath44 and @xmath202 with a number of columns @xmath336 , the complete monotonicity of the function @xmath33 does not ensure that block gauss - type quadrature rules provide bounds on the quantities of the form @xmath335 . in this case , indeed , no information about the sign of the quadrature error can be obtained from the remainder formula for the gauss quadrature rules @xcite .
therefore , we focus on the computation of approximations for the quantities of interest , rather than on bounds .
we propose two different approaches to compute the quantities .
the first one exploits the connection between generalized matrix functions and standard matrix functions described in theorem [ thm : genfun ] , while the second one first approximates the action of a generalized matrix function on @xmath151 vectors and then derives the approximation of the quantities of interest . as a first approach , we propose the use of a pair of block gauss and anti - gauss quadrature rules @xcite based on the nonsymmetric block lanczos algorithm @xcite .
as already pointed out , if we let @xmath290 , it holds that @xmath337 where @xmath338 and @xmath339 . in this case
, there is no equivalent to the polarization identity and thus we work directly with the blocks @xmath340 and @xmath202 , @xmath341 ( the case when @xmath44 and @xmath342 are the initial blocks is similar ) .
let @xmath343 and @xmath344 have all zero entries .
assume moreover that @xmath345 and @xmath346 satisfy @xmath347 .
then the nonsymmetric block lanczos algorithm applied to the matrix @xmath348 is described by the following recursions : @xmath349 @xmath350 . in , @xmath351 and @xmath352 are the qr factorizations of @xmath353 and @xmath354 , respectively , and @xmath355 is a singular value decomposition of the matrix @xmath356 .
the recursion formulas ensures that @xmath357 .
more succinctly , after @xmath255 steps , the nonsymmetric block lanczos algorithm applied to the matrix @xmath348 with initial blocks @xmath358 and @xmath359 yields the decompositions @xmath360 & = \left[\widetilde{z}_1,\dots,\widetilde{z}_\ell\right]j_\ell+\widetilde{z}_{\ell+1}\gamma_\ell \bm{e}_\ell^t,\\ x\left[w_1,\dots , w_\ell\right ] & = \left[w_1,\dots , w_\ell\right]j_\ell^t+w_{\ell+1}\delta_\ell \bm{e}_\ell^t , \end{aligned}\ ] ] where @xmath361 is the matrix @xmath362 and @xmath363 , for @xmath279 are @xmath364 block matrices which contain @xmath128 zero blocks everywhere , except for the @xmath31th block , which coincides with the identity matrix @xmath127
. we remark that if @xmath365 , the use of the symmetric block lanczos algorithm is preferable . in this case , the matrix @xmath361 is symmetric and the decompositions can be written as @xmath366=\left[w_1,\dots , w_\ell\right]j_\ell+w_{\ell+1}\gamma_\ell \bm{e}_\ell^t.\ ] ] the @xmath255-block nonsymmetric gauss quadrature rule @xmath281 can then be expressed as @xmath367 the @xmath306-block anti - gauss quadrature rule @xmath368 is defined as the @xmath306-block quadrature rule such that @xmath369 where @xmath370 , with @xmath371 $ ] , @xmath372\in\mathbb{r}^{k\times n}$ ] , and @xmath373 is the set of polynomials of degree at most @xmath374 ( see @xcite ) . as shown in @xcite , the @xmath306-block nonsymmetric anti - gauss rule can be computed in terms of the matrix @xmath375 as @xmath376 where @xmath377 a pair of block gauss and anti - gauss quadrature rules is not guaranteed to provide upper and lower bounds , not even in the case @xmath378 .
however , suppose that the function @xmath379 can be written as @xmath380 where @xmath381 are the orthonormal polynomials implicitly defined by the scalar lanczos algorithm . in @xcite , the authors show that if the coefficients @xmath382 decay rapidly to zero , then @xmath383 that is , a pair of scalar gauss and anti - gauss rules provides _ estimates _ of upper and lower bounds on the bilinear form of interest .
this result has been extended to the block case in @xcite . in this framework , if we express @xmath230 in terms of orthonormal polynomials , the coefficients in the expansion are @xmath384 matrices . to obtain good entrywise approximations for the quantities of interest
it is necessary that the norm of the coefficients decays rapidly as @xmath255 increases .
this condition is satisfied if @xmath230 is analytic in a simply connected domain @xmath385 enclosing the spectrum of @xmath242 , as long as the boundary @xmath386 is not close to the spectrum @xcite .
if the function @xmath230 satisfies the above conditions , the arithmetic mean @xmath387 between gauss and anti - gauss quadrature rules can be used as an approximation of the matrix - valued expression @xmath388 .
the second approach extends to the block case the approach described in subsection [ ssec : approach3 ] .
assume that the initial block @xmath389 satisfies @xmath390 and that the matrices @xmath391 and @xmath392 are zero matrices .
the following recursions determine the first @xmath255 steps of the block golub
kahan algorithm with starting block @xmath393 : @xmath394 where @xmath395 and @xmath396 are qr factorizations of @xmath353 and @xmath354 , respectively .
after @xmath255 steps , the recursions yield the decompositions @xmath397&=\left[p_1,\dots , p_\ell\right]b_\ell , \\
a^t\left[p_1,\dots , p_\ell\right]&=\left[q_1,\dots , q_\ell\right]b_\ell^t+q_{\ell+1}\gamma_\ell \bm{e}_\ell^t , \end{aligned}\ ] ] where now @xmath398 following the same reasoning as in subsection [ ssec : approach3 ] , when @xmath399 , we can approximate the quantities of interest as @xmath400 f^{\diamond}(b_\ell ) \bm{e}_1 = f_\ell.\ ] ]
in this section we present some numerical results concerning the application of the previously introduced techniques to the computation of centrality and communicability indices in directed networks .
the first set of experiments concerns the computation of the _ total hub communicability _ of nodes , which , for a node @xmath31 , is defined as the following bilinear form : @xmath401_i = { \bf e}_i^t \sinh^\diamond(a ) { \bf 1}\,,\ ] ] where @xmath0 is the adjacency matrix of the digraph .
as shown in @xcite , this quantity can be used to rank how important node @xmath31 is when regarded as a hub " , i.e. , as a broadcaster of information ( analogous quantities rank the nodes in order of their importance as authorities " , i.e. , receivers of information ) .
the second set of experiments concerns the computation of the resolvent - based communicability @xcite between node @xmath31 , playing the role of broadcaster of information , and node @xmath38 , acting as a receiver .
the quantities of interest here have the form @xmath402_{ij}$ ] , where @xmath403 and @xmath404 . in all the tests we apply the approaches previously described and we use as stopping criterion @xmath405 where @xmath406 is a fixed tolerance and @xmath407 represents the approximation to the bilinear form of interest computed at step @xmath255 by the method under study .
our dataset contains the adjacency matrices associated with three real world unweighted and directed networks : , , and @xcite .
the adjacency matrix associated with is @xmath408 and has 7281 nonzeros .
the graph contains information concerning the cross - references in roget s thesaurus .
the adjacency matrix associated with is an @xmath409 matrix with @xmath410 nonzeros . for this network
, there is a connection from node @xmath31 to node @xmath38 if user @xmath31 indicated user @xmath38 as a friend or a foe .
the last network used in the tests , , represents the italian wikipedia .
its adjacency matrix is @xmath411 and has @xmath412 nonzeros , and there is a link from node @xmath31 to node @xmath38 in the graph if page @xmath31 refers to page @xmath38 . in this section
we want to investigate how the three approaches defined for the case of @xmath206 perform when we want to approximate ( [ eq : tc ] ) , the total communicability of nodes in the network . for each network in the dataset , we computed the centralities of ten nodes chosen uniformly at random among all the nodes in the graph .
the results for the tests are presented in tables [ tab : roget]-[tab : itwiki2 ] . the tolerance used in the stopping criterion
is set to @xmath413 .
the tables display the number of iterations required to satisfy the above criterion and the relative error of the computed solution with respect to the exact " value of the bilinear form . the latter has been computed using the full svd for the smallest network , and using a partial svd with a sufficiently large number of terms @xmath414 for the two larger ones . the relative error is denoted by @xmath415 concerning the first approach , since @xmath416 is not completely monotonic , we have used the gauss quadrature rule as an approximation for the quantities of interest , rather than as a lower bound .
.network : , @xmath417 ( @xmath418 ) .
[ cols="^,^,^,^,^,^,^ " , ] tables [ tab : block7]-[tab : block9 ] show the results concerning the approximation of the communicabilities among @xmath151 nodes using the generalized matrix function induced by @xmath419 .
as above , we consider three different values for the parameter @xmath420 . as in the scalar case ,
the method requires fewer iterations to reach a higher accuracy as the value of @xmath420 moves away from @xmath421 .
the two approaches behave again very differently . as before , the results obtained with the second approach are very promising also in view of the fact that we did not make any assumptions on the regularity of the function .
in this paper we have proposed several algorithms for the computation of certain quantities associated with generalized matrix functions .
these techniques are based on gaussian quadrature rules and different variants of the lanczos and golub kahan algorithms . in particular ,
we have investigated three distinct approaches for estimating scalar quantities like @xmath422 , and two block methods for computing matrix - valued expressions like @xmath204 .
the performance of the various approaches has been tested in the context of computations arising in network theory . while not all methods can be expected to always perform well in practice ,
we have identified two approaches ( one scalar - based , the other block - based ) that produce fast and accurate approximations for the type of problems considered in this paper .
francesca arrigo and caterina fenu would like to thank the department of mathematics and computer science of emory university for the hospitality offered in 2015 , when part of this work was completed .
, _ application of anti - gauss quadrature rules in linear algebra _ , applications and computation of orthogonal polynomials , w. gautschi , g. h. golub , and g. opfer , eds . , birkhuser , basel , 1999 , pp .
4156 . , _ preconditioned iterative regularization for ill - posed problems _ , in l.
reichel , a. ruttan , and r. s. varga , eds . , _ numerical linear algebra .
proceedings of the conference in numerical linear algebra and scientific computation , kent , ohio , usa , march 1314 , 1992 _ , de gruyter , berlin and new york , 1993 , pp .
141163 . , _ decay behaviour of functions of skew - symmetric matrices _ , in proceedings of hercma 2005 , 7th hellenic - european conference on computer mathematics and applications , september , 22 - 24 , 2005 , athens , e. a. lipitakis , ed .
, electronic editions lea , athens . | we develop numerical algorithms for the efficient evaluation of quantities associated with generalized matrix functions [ j. b. hawkins and a. ben - israel , _ linear and multilinear algebra _ , 1(2 ) , 1973 , pp .
163171 ] .
our algorithms are based on gaussian quadrature and golub kahan bidiagonalization .
block variants are also investigated .
numerical experiments are performed to illustrate the effectiveness and efficiency of our techniques in computing generalized matrix functions arising in the analysis of networks .
generalized matrix functions , gauss quadrature , golub kahan bidiagonalization , network communicability 65f60 , 15a16 , 05c50 |
current theories of galaxy formation are based on the view that the dominant mass contribution to the universe is in the form of cold dark matter ( dm ) , which clusters under gravity and builds up the backbone of all cosmic structure . together with the discovery of a large dark energy ( de ) component @xcite , this led to the emergence of the concordance @xmath0 cold dark matter ( @xmath0cdm ) cosmogony .
the third component to the energy density in this scenario , the baryons , condense via radiative cooling at the centres of a population of hierarchically merging dm haloes , forming galaxies @xcite .
although the precise physical nature of dm and de is not yet known ( but see * ? ? ?
* for a review of possible dm candidates ) , large - scale predictions of this @xmath0cdm theory show good agreement with a wide range of observations , among them cosmic microwave background ( cmb ) fluctuations @xcite , large - scale clustering of galaxies in the low - redshift universe @xcite , and the redshift @xmath1 power spectrum probed by the lyman-@xmath2 forest @xcite .
it is however crucial to further constrain and test the @xmath0cdm paradigm , especially on the smaller scales of galaxies , which remains a theoretical and observational frontier .
structure formation is a highly non - linear process and only the early stages can be described analytically using linear theory . however , some statistical properties of the evolved dm field , in particular the halo mass function , can also be predicted analytically by combining linear theory and spherical collapse models , through the press - schechter formalism @xcite and subsequent extensions ( e.g. * ? ? ?
these predictions ultimately rely on calibration through more accurate calculations , which typically can only be done numerically .
indeed , while press - schechter theory has been quite successful in describing the abundances of haloes , once computer models reached high precision it became clear that the original analytic treatment was not sufficiently accurate for the quantitative work required in today s era of precision cosmology ( e.g. * ? ? ? * ) .
over the last two decades , numerical simulations have played a key role in guiding our knowledge of structure and galaxy formation .
this is especially evident for the cdm component , where the relevant calculations have reached a high level of sophistication ( e.g. * ? ? ?
* ; * ? ? ?
? * ; * ? ? ?
* ; * ? ? ?
many important insights have been gained from purely collisionless simulations , for example the nearly universal density profiles of dm haloes @xcite , assembly bias @xcite , and the internal structure of haloes on small scales @xcite .
an important reason for today s reliability of cdm predictions is that the initial conditions are unambiguously specified in the @xmath0cdm model , with parameters that are tightly constrained by cmb observations @xcite .
moreover , the computational problem is well - posed and comparatively simple , with equations of motion that for dm involve only gravity .
efficient new algorithms and the rapid growth of computing power over the few last decades have allowed ever more detailed theoretical predictions for the dm distribution .
it is worthwhile , however , to recall that initially such collisionless dm simulations suffered from numerical artifacts , leading to issues like the infamous `` over - merging problem '' , which resulted in featureless and smooth dm haloes primarily due to insufficient mass- , force- and time - resolution @xcite . to understand properties of the observable universe one needs to relate the dark matter distribution to that of the baryonic material
this can be done in a number of different ways : ( i ) using so - called semi - analytical modelling of galaxy formation ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or the closely related approach of halo occupation distributions @xcite , or ( ii ) using direct hydrodynamical simulations ( e.g * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . approach ( i ) allows a fast exploration of a large parameter space of the underlying coarse description of galaxy formation physics , whereas ( ii ) makes possible the calculation of a self - consistent model that requires far fewer assumptions about the gas dynamics .
nevertheless , both methods share the common problem that complicated physics on very small scales , like star formation and associated feedback processes , need to be accounted for in a rather crude way .
this propagates into uncertainties in the interpretation of simulation results for the hydrodynamic sector .
one possible strategy to get better control over this problem is to calibrate sub - resolution treatments in semi - analytic models and hydrodynamical simulations observationally ( e.g. * ? ? ?
* ; * ? ? ?
in addition , numerical artifacts in hydrodynamical simulation techniques need to be better understood , hopefully allowing one to reduce or completely eliminate them .
state - of - the - art cosmological hydrodynamical simulations include a prescription for star formation ( e.g. * ? ? ?
* ) , radiative cooling @xcite , chemical enrichment ( e.g. * ? ? ?
* ) and various feedback processes relevant for low- and high - mass systems ( e.g. * ? ? ?
* ; * ? ? ?
furthermore , some simulation codes can also include magnetic fields @xcite , radiative transfer @xcite , cosmic ray physics , thermal conduction @xcite , or black hole physics @xcite .
hydrodynamical simulations have been successfully used to study the lyman-@xmath2-forest ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and the detailed physics of the intracluster medium , where cluster scaling relations were obtained that agreed with x - ray observations @xcite .
there have also been numerous studies that focused on the formation of a representative galaxy population ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
while some successes have been achieved here as well , a general finding of these simulations is that it appears to be difficult to reproduce the observed shallow slope of the faint - end of the galaxy luminosity function and the observed morphological mix of galaxies , with its high abundance of disk - like galaxies .
although there has been some considerable progress over the last few years on the latter issue ( e.g. * ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , one of the outstanding problems is to produce a statistical sample of galaxies that agrees reasonably well with observations and reproduces the hubble sequence .
but modifications in the detailed feedback and star formation prescriptions proved to be important and successful in forming more realistic late type spiral galaxies .
we note that cosmological hydrodynamic simulations have also made it clear that a proper inclusion of the baryonic physics can lead to interesting back reactions on the dm distribution @xcite , potentially eliminating or reducing the central dark matter cusp @xcite , or forming a dark disk @xcite .
it is hence clear that it is ultimately not sufficient to work with dark - matter only simulations .
most hydrodynamical simulations of galaxy formation carried out thus far have used the smoothed particle hydrodynamics ( sph ) technique , where the gas is discretised into a set of particles for which appropriate equations of motion can be derived .
the sph method is well - suited for cosmological applications owing to its pseudo - lagrangian character , which automatically brings resolution elements to regions where they are needed most , i.e. collapsing objects like clusters and galaxies .
also , the conservation properties of sph in terms of simultaneous conservation of energy , momentum , mass , entropy and angular momentum are excellent .
a further convenient feature is that a particle - based gravity solver ( needed for the dm component anyway ) can be readily applied to sph .
these properties have made the method also popular for applications other than cosmic structure growth , for example for isolated merger simulations . however
, different methods for solving the equations of hydrodynamics in a cosmological context have been employed as well , most of which are descendants of classic eulerian methods @xcite on regular meshes .
the fixed cartesian grids originally available in the corresponding codes are insufficient to capture the large dynamic ranges encountered in galaxy formation , but finite volume schemes incorporated in modern adaptive mesh refinement ( amr ) codes can alleviate this problem .
interestingly , these mesh - based methods have led in some cases to significantly different results compared with sph .
a prominent example of these discrepancies is highlighted in the santa barbara cluster comparison project @xcite , where it was found that the entropy profiles of clusters are significantly and systematically different between amr and sph codes , the former yielding a large entropy core which is absent in the sph calculations .
only recently has some progress been made in identifying the cause for this difference @xcite , but the issue is still not fully understood ; hence we return to it in one of our companion papers @xcite .
it thus appears that there are at least two important challenges in the field of cosmological hydrodynamics simulations .
one lies in an adequate treatment of all relevant physics of galaxy formation , in particular with respect to the reliability of the parameterisations of sub - resolution physics , and the other having to do with the accuracy and efficiency of the underlying hydrodynamics solver .
it is of course crucial to strive for a more faithful treatment of the physics in the simulations , but it should also be evident that a proper modelling of the physics relies on a correct solution of the basic hydrodynamical equations in the first place .
it is therefore problematic to tune the sub - resolution physics without first verifying in detail the quality of the hydro solver , because this risks incorrectly absorbing deficiencies of a numerical technique into distorted physics models . here
, in this initial study we shall primarily be concerned with an assessment of the extent to which uncertainties in numerical methodology are reflected in the predicted galaxy properties , for a fixed physics model . in future works
, we will explore the consequences of various treatments of feedback effects , which are known to be crucial for the galaxy formation problem @xcite .
we remark that it is not obvious whether sph or amr is more accurate in all simulation regimes , as both methods are known to have shortcomings .
for example , sph suffers from relatively poor shock resolution , noise on the scale of the smoothing kernel , and low - order accuracy for the treatment of contact discontinuities .
furthermore , some hydrodynamic instabilities like the kelvin - helmholtz instability can be suppressed in sph @xcite .
recently , @xcite also showed that conventional implementations of sph do not properly properly subsonic turbulence , which is generically present in cosmological haloes .
astrophysical eulerian methods on the other hand ( usually realised as amr finite - volume schemes ) may suffer from over - mixing due to advection errors in the presence of bulk flows .
one of the principal differences between sph and amr lies in their handling of mixing at the level of individual fluid elements .
this is absent in sph by construction ( unless added in crudely by hand somehow ) , while in amr it occurs implicitly through the averaging of the evolved riemann solutions over the scale of the grid - cells at the end of each timestep .
it is not always clear which scheme yields a more accurate result @xcite .
another issue with classical amr codes is that their handling of the discrete equations of motion can lead to errors with the fluid is moving rapidly across the mesh ( * ? ? ?
* hereafter s10 ) .
because the truncation error of eulerian codes depends on the fluid velocity relative to the grid , the results can thus be sensitive to the presence of bulk velocities ( see also * ? ? ?
* ; * ? ? ?
finally , a further challenge for amr schemes in cosmic structure formation arises from the need to accurately follow the gravitational growth of even very small structures .
the mesh - based poisson solvers typically employed by amr codes have been shown to lack sufficient small - scale force accuracy and to produce too few low mass haloes potentially corrupting the solution on even well - resolved scales .
while this can be addressed with more aggressive refinement strategies , the discontinuous changes in gravity resolution brought about by such refinements introduce non - hamiltonian perturbations into the dark matter dynamics which are in principle undesirable .
recently , s10 introduced a new moving - mesh approach as embodied in the arepo code .
the principal goal is similar to the earlier implementations of @xcite and @xcite , but these moving - mesh algorithms suffered from grid distortions which limited their applicability .
arepo does not use coordinate transformations like previous moving mesh codes in cosmology , but instead employs an unstructured voronoi tessellation of the computational domain .
the mesh - generating points of this tessellation are allowed to move freely , offering significant flexibility for representing the geometry of the flow . as discussed in detail in s10
, this technique avoids several of the weaknesses of sph and amr schemes while it retains their most important advantages .
for example , if the mesh motion is tied to the gas flow , the results are galilean - invariant ( like in sph ) , while at the same time a high accuracy for shocks and contact discontinuities is achieved ( like in eulerian schemes ) .
the aim of this paper is to compare the novel cosmological hydrodynamics code arepo with the widely used sph code gadget @xcite . in this first paper of a series
we give an overview of the numerical techniques used for our galaxy formation simulations , introduce our simulation set , present an analysis of the global baryon characteristics , and evaluate the performance and efficiency of the new simulation scheme . in one companion paper (
* hereafter paper ii ) , we discuss the properties and statistics of individual galaxies and haloes in the simulations . a further companion paper ( * ? ?
* hereafter paper iii ) analyses idealised test problems to highlight various differences of the two simulation schemes and to elucidate how they affect galaxy formation .
the outline of this paper is as follows . in section 2
, we provide a discussion of the initial conditions and numerical details which are particularly important for the goal of our comparison project .
there we also discuss some modifications of the arepo code that were adopted during the course of this work .
section 3 presents first results and discusses the large - scale structure and global statistics of the baryon content in the simulations at different resolutions .
we turn to an analysis of the origin of the cooling difference we find between the codes in section 4 . in section 5
we discuss some generic issues and problems of sph .
finally , we give a summary of our results and our conclusions in section 6 . in an appendix
, we provide data on the mesh geometry , analyse the performance of the codes in different parts of the calculations and present scaling tests .
our gadget and arepo simulations follow the same physics , consisting of a collisionless dark matter fluid and an ideal baryonic gas .
both components act as sources of gravity and are evolved in a newtonian approximation on top of an expanding friedman - lemaitre background model .
we describe the gas dynamics with the ordinary inviscid euler equations , augmented with additional terms that account for `` extra physics '' related to radiative processes and star formation .
in particular , we account for optically thin radiative cooling of a primordial mixture of helium and hydrogen , with a hydrogen mass fraction of @xmath3 .
we also assume a spatially uniform , time - dependent ionising uv background with an amplitude and time evolution according to @xcite . as a result of cooling
, gas can collapse to high density and turn into stars .
both simulation codes describe this process with the simple star formation and supernova feedback model introduced in @xcite , which has been used in many sph simulations in recent years . in this approach ,
every sufficiently dense gas particle / cell is treated as a representative region of the interstellar medium ( ism ) and is assumed to exhibit a multiphase structure consisting of cold and hot gas in pressure equilibrium .
stars form out of the cold gas and return energy back into the medium though supernova explosions .
this energy feedback pressurises the multiphase medium , the effect of which is described in terms of an effective equation of state that is imposed onto the gas once it has overcome the density threshold for star formation .
the value of the density threshold we use in this work is @xmath4 , set such that isolated disk galaxies at @xmath5 are characterised with a relation between disk surface gas density and star formation density that agrees with the observed kennicutt law @xcite . to reach these densities gas needs to fall into dark matter haloes and dissipate energy via radiative cooling to eventually form galaxies @xcite .
once the gas density is higher than the threshold value , particles / cells are eligible to form stars and can be converted into collisionless stellar particles .
we treat this conversion as a stochastic process that generates one generation of stellar particles per hydrodynamic resolution element ( sph particle or voronoi cell , respectively ) at an average rate equal to the star formation rate predicted by the subresolution multi - phase model .
we stress again that for the purposes of our study it is crucial to have an identical implementation of this extra physics in both codes in order to allow a straightforward comparison of the two hydrodynamical schemes .
fortunately , the quasi - lagrangian nature of arepo typically allows an easy adoption of methods originally developed for particle - based schemes like sph .
we could hence adopt most of the `` extra physics '' implemented in gadget in a largely unmodified form in arepo , helping to ensure that possible implementation differences for this physics do not affect our conclusions .
we also emphasise that in our present context it is preferable to employ a relatively simple subresolution model like that of @xcite in order to facilitate a clear assessment of the differences between hydro solvers . in particular , for now we have not included the effects of strong feedback capable of inducing galactic winds and outflows .
we are thus not attempting to solve the galaxy formation problem at this stage , but are instead concerned with identifying and understanding issues related to the treatment of the hydrodynamics . in the following two subsections we highlight the most relevant features and parameter settings of the codes we used , as well as any particular changes we made for this project .
we note that full details are presented in substantial depth in the corresponding code papers , so in the interest of brevity we here restrict ourselves only to those aspects directly pertinent to our study .
we are also aided by the fact that the arepo and gadget codes share the same gravity solver , something that was not the case in previous comparison studies of sph and amr ( e.g. * ? ? ?
consequently , the evolution of the collisionless dm component is identical in both codes in the limit of vanishing hydrodynamical forces .
this similar handling of self - gravity is important for isolating differences caused primarily by the hydro solvers .
gadget is a widely used and well - tested sph - code which employs a formulation of sph that simultaneously conserves energy and entropy despite the use of fully adaptive smoothing lengths @xcite .
the gravity calculation is split into short- and long - range components , where short - range forces are calculated with an hierarchical oct - tree algorithm @xcite and the long - range forces are evaluated with a particle mesh ( pm ) method ( e.g. * ? ? ?
our gadget runs are based on a default setting of the numerical sph parameters , using 32 neighbours in smoothed estimates and an artificial viscosity parameter of @xmath6 , combined with balsara s switch @xcite to reduce the viscosity in the presence of strong shear .
arepo is a second - order accurate finite volume method that solves the euler equations using piece - wise linear reconstruction and a calculation of hydrodynamical fluxes at each cell interface with an exact riemann solver .
the basic solution strategy of the code is that of the well - known muscl - hancock scheme .
what makes arepo unusual is that it employs an unstructured mesh based on a voronoi tessellation using a set of mesh - generating points .
furthermore , the mesh is allowed to move freely as a function of time . in our runs
, we tie the motion of the mesh - generating points to the hydrodynamical flow , which gives the scheme a quasi - lagrangian character and makes the mesh automatically adaptive . as s10 demonstrated , such an approach offers a number of advantages compared with ordinary eulerian or lagrangian codes , including a reduction of numerical diffusivity and advection errors and an improved handling of contact discontinuities . in the following ,
we discuss some details and modifications of arepo relevant for our study . *
( i ) mesh regularisation : * it is desirable to have a voronoi mesh geometry where the geometric centre of each voronoi cell lies reasonably close to the mesh - generating point of that cell .
this reduces the size of errors in the linear reconstruction step of the muscl - hancock scheme and also limits the rate at which mesh faces turn their orientation during the mesh motion .
arepo therefore incorporates a method to regularise the mesh , as described in s10 .
this scheme uses a modified lloyd algorithm to drift the mesh a bit towards its centroidal configuration at each time step . to this
end the motion of the mesh - generating points contains an additional velocity component for certain cells , designed to move a mesh - generating point closer towards the geometric centre of its cell .
the default setup of arepo uses the spatial offset between the actual location of a vertex and its cell s geometric centre in units of the cell s size to decide whether or not this corrective motion should be added to the cell s vertex velocity .
however , a number of test simulations revealed that this regularisation scheme can introduce some unwanted side effects .
specifically , systematic differences in the average mass of gaseous cells can develop between the star - forming gas and the lower density material in the surrounding halo , with the former tending to become systematically more massive than the latter .
that such a tendency exists in principle is to be expected because the lloyd algorithm would ultimately converge to a homogeneous voronoi mesh with cells of equal volume if it applied to every cell centre with full intensity all the time .
the mesh regularisation motions in the default scheme have hence the tendency to move the mesh away from the densest regions ( but leaving the gas there in arepo , the gas motion and the mesh motion are independent ) , such that the mass per cell will tend to increase in the densest regions .
this is somewhat undesirable because ideally we want to have the highest mass resolution in the central parts of galaxies and not in the surrounding low density gas .
there are at least two solutions to this problem ( besides disabling the regularisation scheme altogether ) .
one is to simply make the settings of the standard regularisation less aggressive , thereby reducing the systematic effects described above .
another possibility is to change the criterion that decides when such a mesh correction motion should be applied .
we have found that the simple criterion originally implemented in arepo , based on the displacement of a mesh - generating point from the geometric centre of its cell , is prone to invoke mesh regularisation drifts even when they are not needed .
this happens especially in regions with strong density gradients , where such displacements naturally occur even if the mesh geometry is acceptable .
as an alternative criterion we have therefore considered the maximum opening angle under which a cell face is seen from the mesh - generating point .
this identifies problematic cell geometries more directly , which are in fact those where a point lies close to a wall , or equivalently , where this opening angle is large .
we find that this opening angle criterion is more tolerant to rapid changes in spatial resolution , and is hence less prone to triggering unwanted mesh - correction motions in the presence of large density gradients .
because of this it virtually eliminates the mass segregation effect mentioned above . for definiteness , in the simulations presented here , we invoke mesh - regularisation motions if the maximum face angle of a voronoi cell ( defined as the maximum of @xmath7 , where @xmath8 is the area of a face and @xmath9 its distance to the vertex ) is larger than @xmath10 . in this case
we add a velocity component up to half the sound speed of the corresponding cell to the vertex velocity , which moves the mesh - generating point closer towards the geometric centre of the voronoi cell .
ideally , this scheme should then guarantee reasonably regular cells where the bulk of all cells will have a maximum face angle less than @xmath10 .
our tests confirm that this is the case , while at the same time the mass bias between star forming high density cells and the lower density cells in the surrounding halo is significantly reduced .
the face angle scheme applies the mesh - correction motions more rarely than the original approach but still frequently enough to maintain a regular mesh . in the appendix
, we discuss some measurements of the mesh statistics , which also show that this new regularisation scheme still ensures that the offsets between mesh - generating points and the geometric centres of the corresponding cells remain small , as is desired to minimise errors in the linear reconstruction step . *
( ii ) refinement and de - refinement : * arepo is quasi - lagrangian in the sense that the vertices of the voronoi mesh follow the flow of the fluid such that an approximately constant mass resolution is automatically achieved .
the code is not purely lagrangian , however , because mass can be exchanged between cells in a manner consistent with the equations of motion , in order to prevent the mesh from becoming highly distorted .
moreover , the code can also make use of re- and derefinement of individual cells in order to improve the local resolution beyond that offered by the dynamical motion of the mesh or to improve efficiency when high resolution is no longer required in certain regions .
the method employed by arepo is more general than that used by standard amr methods , which typically employ subgrids to refine a certain section of the parent grid , thereby creating a hierarchy of overlapping grid patches . in arepo , we are not required to use subgrids , but can split or merge individual cells , such that a single global mesh with a smooth transition from low- to high - resolution parts results .
the criteria for when a change of the local resolution should be introduced are very flexible , just as in the amr technique . indeed , because the initial distribution of mesh generating points and the manner in which they are updated in time are arbitrary , arepo offers the capability of running in an amr mode . in our cosmological galaxy formation simulations
it is desirable to maintain a roughly constant mass resolution .
this is because in runs with star formation whole gas cells can turn into collisionless stellar particles , and it is best to create them with approximately the same mass to avoid potential two - body heating effects among these star particles . a relatively homogeneous mass resolution in the gas phase is also warranted to guarantee a more proper comparison to the sph runs , where a fixed mass resolution is present by construction .
( we note , however , that this fixed mass resolution in sph is a consequence of this method s inability to correctly represent the flow on small scales , as we discuss in section 5.3 . )
we have therefore introduced a new re-/derefinement scheme in arepo that ensures that the mass resolution in the gas and the spawned star particles never deviates too much from the value of the sph particle mass in the corresponding gadget simulation .
specifically , we `` force '' cells to have a mass in the range @xmath11 , where we set the target gas mass @xmath12 equal to the mean cell mass assuming a homogeneous gas distribution in the simulation volume with equal volume cells .
this corresponds then exactly to the mass of an individual sph particle in the gadget calculation at the same particle / cell number .
we allow only quite `` roundish '' cells to be refined , specifically we only refine a cell if the maximum face angle is smaller than @xmath13 .
this protects against introducing large local errors by splitting very irregular cells ( which are usually cells that have just been split in the previous timestep ) . as a result of this constraint
our simulations may contain at any given time a few cells which are more massive than @xmath14 .
we have also done runs where only cells with a sufficiently shallow temperature gradient across them were allowed to be derefined , but this did not make any significant difference to our results .
we also slightly modified the star formation implementation to take this re-/derefinement scheme into account .
specifically , if a star particle is formed in a cell with mass less than @xmath15 , it inherits the full mass and the cell is removed .
otherwise , a star particle is created with a mass equal to @xmath16 , leaving the rest of the mass in the ( surviving ) cell .
cells that do not contain at least @xmath17 do not produce stellar particles , which prevents very low mass collisionless particles from being formed . we note that such low mass cells are exceedingly rare , because cells with less than @xmath18 are normally all removed during the derefinement process .
however , since two neighbouring cells will never be dissolved in the same timestep ( see the discussion in s10 ) , a cell can in rare cases briefly scatter below @xmath19 and potentially also attempt to spawn a stellar particle during this time .
this modification of the star formation implementation guarantees that star particle masses are bounded by the interval @xmath20 .
our gadget and arepo simulations both use only one generation of stellar particles , which implies that the total number of baryonic resolution elements in the gadget simulations is strictly conserved , while in arepo it can fluctuate slightly around the initial value . *
( iii ) dual entropy formalism : * as discussed in s10 , finite volume hydro solvers can suffer from spurious heating in highly supersonic , cold parts of the flow , where the kinetic energy dominates over the thermal energy . especially at high redshifts in cosmological simulations
, this can lead to an incorrect temperature evolution in low - density gas as a result of discretisation errors in the advection of the kinetic energy .
owing to its quasi - lagrangian character arepo is somewhat less prone to this effect than fixed mesh techniques , but it can still be present at some level . to circumvent this problem , arepo can optionally force the entropy to be conserved instead of the energy during a timestep . in this case , no spurious heating will be introduced .
however , entropy is conserved only outside of shocks in adiabatic parts of the flow , so a criterion is required to decide on a cell - by - cell basis whether energy or entropy conservation should be given precedence for a cell s evolution over the current timestep .
as s10 showed , this dual entropy formalism is especially useful for non - radiative simulations , because here a spurious heating can not be compensated by radiative cooling .
since our simulations include radiative cooling due to a primordial gas mixture we expect this to be less problematic for our applications .
indeed , extensive test simulations showed that our results are not sensitive to whether or not the dual entropy formalism is used .
we have therefore decided to exclusively use strict energy conservation in all our production calculations , without employing the dual entropy formalism . *
( iv ) gravitational softening length : * gadget uses a global gravitational softening length for collisional and collisionless particles , i.e. for gas and stellar / dark matter particles .
arepo also uses a global gravitational softening length for collisionless particles ( which we set equal to that used in our gadget simulations ) .
however , as the code is a finite volume method which uses gas cells of variable size to represent the fluid , we have used individual plummer - equivalent gravitational softening lengths according to @xmath21 for each gaseous cell , where @xmath22 is the cell s fiducial radius assuming the cell volume is spread over a sphere , and @xmath23 is an overall scaling factor . in addition , we imposed a lower floor on the gravitational softening lengths equal to that of the sph particles in the corresponding gadget simulation .
this was done in order to safeguard against a potentially higher gravitational resolution in arepo in the densest gas , which may have distorted our comparison .
however we note that we do not expect any significant difference due to the slightly different treatment of the gravitational softening length of the gas even if such a floor was not used .
indeed , as described in appendix b , we have repeated lower resolution arepo simulations with a fixed softening length for all gas cells and obtained nearly identical results . also , we have repeated simulations with adaptive gravitational softening without a lower floor and again obtained equivalent results .
we note that the presence of an effective equation of state for highly dense gas is ultimately responsible for this insensitivity . without it
, we would get fragmentation of very dense gas where the adaptive softening could make a difference @xcite . [
cols="^,^,^,^,^,^,^,^",options="header " , ] in any comparison between codes as different as gadget and arepo , a number of different benchmarks can be used to gauge their relative performance , and various criteria can be defined to align runs that are deemed comparable with each other . in our study , we have chosen to compare gadget and arepo at matching mass resolution , using the same initial number of baryonic and dark matter resolution elements .
this provides a well - defined comparison strategy and is advantageous for a number of reasons : \(1 ) first , we can use the same initial conditions for the simulations with both codes .
this would not be possible if we performed the runs with different initial mass resolutions .
( 2 ) second , our choice makes it straightforward to relate the same objects between the different runs ( and they are resolved in dark matter in the same way ) .
( 3 ) third , as we discuss below , for the same initial mass resolution , the cpu time requirements for gadget and arepo are similar , at least for cosmological simulations , with the moving mesh runs being only some @xmath24 more costly . in many cases ,
the limiting factor for the simulation size that can be achieved is the amount of computing time required , so our choice effectively enables a comparison for fixed computational resources . of course , in principle one may also attempt a comparison at equivalent spatial resolution in the gas between gadget and arepo , or at the same accuracy in the solution obtained .
we have not attempted this for the following reasons .
first , while the spatial resolution that can be achieved for a given number of resolution elements is expected to be worse in sph compared to arepo , the degree to which this matters is likely very problem - dependent .
for example , for our treatment of star formation , the mass resolution is much more important than a precise treatment of fluid discontinuities . aligning the spatial resolutions such that discontinuities are broadened over the same scale in both codes would lead to drastically different mass resolutions in the star - forming phases and hence likely to large systematic resolution effects in the smallest galaxies .
second , and even more important in our view , is that the convergence properties of sph are not well understood , making a comparison at fixed accuracy a highly problematic undertaking from the start . as we discuss at length in section 5 , formal convergence in sph
ultimately requires that the number of neighbouring particles used in smoothed estimates around a particular location be increased as the total number of particles is made larger .
this has typically not been done in cosmological applications , and is also in practice not readily possible due to clumping instabilities ( e.g. * ? ? ?
moreover , while sph is often described as being `` lagrangian , '' this is not formally correct because individual fluid elements , as represented by sph particles , are not allowed to deform arbitrarily in response to e.g. shearing motions , as we illustrate in section 5 .
sph should , therefore , properly be referred to as `` pseudo - lagrangian . ''
the errors that this feature of sph entails depend on the detailed properties of the flow , in addition to the local spatial resolution , and consequently can not be assessed in general . in its most basic formulation ,
sph may simply lack a formal convergence condition , making it unclear how a comparison at fixed accuracy should be defined .
finally , in our study , we have chosen a particular implementation of `` standard sph '' , as implemented by the gadget code . moreover , by choosing gadget for our sph simulations , we are able to use the same gravity solver and the same physics in our comparisons between sph and arepo , which might not be possible otherwise .
this enables us to isolate differences in runs that owe primarily to the differences in the hydro solvers between gadget and arepo .
as we discuss in section 5 , various modifications have been proposed to the basic structure of sph in order to improve its reliability under some circumstances .
it is , of course , possible that we would have obtained somewhat different results had we employed such formulations of sph .
however , we believe that many limitations of sph are generic and are not specific to gadget . indeed , as we discuss in section 5 , it is difficult to see how sph could be modified to entirely eliminate , for example , the sources of error that owe to its pseudo - lagrangian nature without radical modifications to the underlying algorithm .
we also stress again that the primary goal of our comparison is not to arrive at the best fit to observational data , or to achieve the highest possible resolution for single galaxy models . while we acknowledge that important strides have been made recently in studying disk formation numerically using zoom - in procedures in cosmological sph simulations ( e.g. * ? ? ?
* ; * ? ? ?
* ) , we do not think these successes provide much evidence for the numerical reliability of sph , although this misconception appears widespread .
our ultimate aim is not merely to model single objects , but entire populations of galaxies so that we can statistically constrain uncertain processes associated with star formation and feedback by using , e.g. , observations of the distribution of galaxies with respect to their hubble type .
this requires the identification of accurate and computationally efficient techniques that are best suitable for the task .
in particular , our strategy is designed to detect systematic inaccuracies in current simulation techniques that may easily be concealed by the adjustment of heuristic feedback parameters to match a certain observation , as it is commonly done , while at the same time these effects may have serious consequences in other places . for our simulation
set we adopt the cosmological parameters @xmath25 , @xmath26 , @xmath27 , @xmath28 , @xmath29 and @xmath30 ( @xmath31 ) .
these parameters are consistent with the most recent wmap-7 measurements @xcite as well as with a host of other cosmological constraints .
we create a realisation of this cosmology in a periodic box with a sidelength @xmath32 , at three different mass resolutions corresponding to @xmath33 , @xmath34 and @xmath35 particles / cells .
a comoving gravitational softening length of @xmath36 , @xmath37 or @xmath38 , respectively , is used for the dark matter , and for the gas particles in the gadget runs . for arepo , we start initially with the same number of cells as there are particles in the corresponding sph run , but we use an adaptive gravitational softening with a floor as described earlier . note that due to re- and derefinement the number of baryonic resolution elements ( cells plus star particles ) is not strictly conserved during the course of a simulation but can fluctuate slightly around the initial number .
initial conditions were generated at @xmath39 based on the power spectrum fit of @xcite , with gas particles / cells added to the initial conditions by splitting each original particle into a dark matter and gas particle / cell pair , displacing them with respect to each other such that two interleaved grids are formed , keeping the centre - of - mass of each pair fixed .
all our simulations have been evolved until redshift @xmath5 even though the fundamental mode of our small box becomes mildly non - linear at low redshift . however , our primary goal here is not to obtain quantitatively accurate large - scale statistics , but rather to compare the properties of galaxies formed by two different hydrodynamical schemes .
as we start all simulations with the same phases , this relative comparison is unaffected by box - size effects down to @xmath5 . to easily refer to the different runs in our simulation suite we use the following naming convention .
a_l20nx denotes arepo runs ( indicated by the leading ` a ' ) , where @xmath40 , @xmath41 , or @xmath42 encodes the numerical resolution and the tag ` l20 ' stands for the box size of @xmath43 .
likewise , g_l20nx refers to our gadget runs ( indicated by the ` g ' ) . in table
[ table : simulations ] , we provide an overview of our simulation suite and list its most important parameters .
to obtain a first impression of the simulations , it is instructive to examine maps of the density and temperature fields . in fig . [
fig : projections ] , we show projections of the gas density and temperature distributions for our highest resolution simulations a_l20n512 and g_l20n512 at @xmath44 . the left panel gives the arepo results and the right panel those obtained with gadget . in both cases , the image on top depicts a large part of the full box , while the zoom images below show two successive zooms onto a disk galaxy ( as indicated by the white squares ) . in these enlarged images , the left half depicts the temperature field while the right half gives the density field .
all slices have a thickness of @xmath45 and their extent in the image plane is specified by the scale - bars included in the images .
note that we did not centre the two image sequences individually onto the galaxies , so the zoom sequence also illustrates the typical magnitude of coordinate offsets that can develop between the different simulations . inspection of the maps in fig .
[ fig : projections ] demonstrates that both codes produce essentially identical results on large scales .
this is reasonable since we do not expect large - scale changes induced by a different hydrodynamical scheme .
however , already at the intermediate zoom level one can identify some significant differences .
the distribution of hot gas is clearly quite different , with gadget showing a more extended hot atmosphere compared to arepo . on still smaller scales ,
the differences between the two hydrodynamical codes become even more pronounced : whereas arepo produces an extended disk , gadget forms a significantly smaller and clumpier disk .
the difference can be best appreciated in the temperature map , where the cold gas disk stands out clearly as a large blue region in the moving mesh calculation .
arepo forms a spiral disk with a central bar , which is visible in the density map . for a more detailed analysis of the structural properties of this galaxy we refer the reader to paper ii of this series .
we note that although some spiral features are visible in the disk , these are partly seeded by poisson noise in the potential due to the limited number of dm particles in the halo ( see * ? ? ?
* for details of this process ) , and have hence only limited physical significance . in order to demonstrate that differences like those seen in fig .
[ fig : projections ] do not appear only in unusual cases , we show in fig .
[ fig : collage_gas_z2 ] gas density projections of the central galaxies of the 24 most massive haloes at @xmath44 of a_l20n512 ( top panels ) , and g_l20n512 ( bottom panels ) , along random projection directions .
we did not impose any constraint on the galaxy selection other than halo mass ; as a result , we can also have galaxies in our sample that are disturbed or undergo a merger , such as the system labelled ` galaxy - id 5 ' .
the object ` galaxy - id 8 ' of fig .
[ fig : collage_gas_z2 ] corresponds to the galaxy shown in fig .
[ fig : projections ] , but happens to be oriented edge - on here .
the random orientations of the galaxy set give an impression of how a population of galaxies would look like in these simulations . clearly , the gas distributions of most of the galaxies in fig .
[ fig : collage_gas_z2 ] are more extended and more reminiscent of spiral galaxies in arepo than in gadget .
inspection of animated sequences from the arepo simulations reveals a population of interacting and merging systems .
many of them show thin bridges and tails , as expected if the galaxies are rotationally supported disks ( e.g. * ? ? ?
* ; * ? ? ?
* ) and @xmath44 ) and high - resolution images are available for download at the website http://www.cfa.harvard.edu/itc/research/movingmeshcosmology . ] .
the differences in the gas properties are also reflected in the corresponding stellar distributions of these galaxies , which we show in fig .
[ fig : collage_stars_z2 ] .
although the stellar distributions are more similar between the two numerical schemes , in most cases one still notices that arepo s disks are slightly larger and appear typically more disky , an impression that is also supported by the quantitative analysis of galaxy properties that we present in paper ii
. furthermore , the stellar populations of the arepo galaxies are bluer , i.e. younger , than those found in the gadget run . in fig .
[ fig : collage_gas_z0 ] and fig .
[ fig : collage_stars_z0 ] we show face - on projections of 32 other galaxies at @xmath5 with milky way - like dm halo masses in the mass range from @xmath46 to @xmath47 .
the gas disk structure shown in fig .
[ fig : collage_gas_z0 ] follows the trend already seen at @xmath44 , and demonstrates that arepo produces significantly larger disks at all times and at all halo masses compared to the gadget simulations .
[ fig : collage_stars_z0 ] demonstrates that the stellar disks are also slightly larger , but not as significant as the gas disks .
but similar to the @xmath44 result , the arepo stellar disk populations are typically younger than those found in gadget . a self - evident conclusion from these visual comparisons of figs . [ fig : collage_gas_z2 ] , [ fig : collage_stars_z2 ] and figs .
[ fig : collage_gas_z0 ] , [ fig : collage_stars_z0 ] is that the morphology of forming galaxies in cosmological hydrodynamics simulations can be strongly affected by the underlying hydrodynamics solver .
sph has significantly more problems producing disk - like galaxies than the moving mesh approach for a similar computational cost .
note however that we do not expect our simulations to produce a completely realistic @xmath44 galaxy population , because we here refrained from including strong feedback capable of producing galactic winds and outflows .
however , since both simulations followed exactly the same physics , we expect this qualitative difference to be present in simulations with stronger feedback as well .
as we emphasised earlier , arepo and gadget use the same gravity solver based on a high - resolution tree - pm scheme . because the gravitational field is largely dominated by the collisionless dm component
, we expect that the dm distribution in both simulations should hence be similar , at least on scales where baryonic effects do not change the structure of dm haloes .
we explicitly verify this expectation in the left column of fig .
[ fig : fof_mf ] , where we compare the dm halo mass function of friends - of - friends ( fof ) groups with a linking length of @xmath48 of the mean particle separation down to a particle number limit of 32 particles at @xmath5 and @xmath44 .
the codes show excellent agreement in the dm halo mass functions at both redshifts . also , the convergence for the different resolutions is in both cases very good .
we note that previous comparisons of hydrodynamical mesh codes and high resolution tree codes have shown a deficit of low - mass haloes in the mesh - based hydrodynamical codes , an effect that has its origin in their amr - based gravity solver , which typically does not refine small dm haloes sufficiently early .
as evidenced here , arepo does not have this problem .
the other panels of fig .
[ fig : fof_mf ] count haloes in terms of their baryonic content , separately for gas and stars . in order to relate baryons to dark matter haloes ,
we determine for each stellar and gaseous cell / particle the closest dark matter particle , and then associate the baryonic mass element with the fof group this dm particle resides in ( see * ? ? ?
* for more details ) . the middle column of fig .
[ fig : fof_mf ] shows the cumulative halo gas mass functions at redshifts @xmath44 and @xmath5 .
while both codes converge equally well for the gas mass functions as a function of resolution , they do not agree on the result . at @xmath44
, arepo has more objects of a given gas mass than gadget over nearly the full range of halo masses .
this behaviour changes towards lower redshifts , where fewer very gas - rich objects are found in arepo
. we will show below that this has to do with more efficient cooling and a higher star formation rate in the mesh code in massive haloes at late times . the right column in fig .
[ fig : fof_mf ] gives the mass functions for the stellar component of haloes , and here the situation is slightly different . at @xmath44 , we find fewer objects with low stellar mass in arepo compared to gadget , but the situation reverses at the high mass end where we find arepo has more haloes at a given stellar mass
. this behaviour changes somewhat towards @xmath5 , where the low mass end agrees now very well between the two codes .
however , there are still more stellar systems of high mass in arepo , and this difference becomes more pronounced and extends over a larger range of masses .
again , this can be attributed to the more efficient star formation at late times in massive systems in arepo that we analyse in detail below .
since the baryonic mass contributes only a small amount to the total mass of the entire box , these differences in the baryonic mass functions are neither imprinted in the dm halo mass functions nor visible in the large - scale structure shown in fig .
[ fig : projections ] .
because we use two completely different hydro solvers , we may expect some deviations in the breakdown of various gas phases .
a first simple way to obtain an overview of the global state of the gas in the simulation volume is the construction of density temperature phase - space diagrams .
we show in fig .
[ fig : rho_t ] mass - weighted @xmath49-@xmath50 phase - space diagrams of the simulations at the highest resolution for @xmath44 ( top row ) and @xmath5 ( bottom row ) .
the left and right columns show results for arepo and gadget , respectively .
we can readily identify a number of well - known features in the gas phase - space @xcite , which show up in both runs .
the narrow ridge of gas with an upward slope and @xmath51 and @xmath52 consists of low density , highly photo - ionised gas in the intergalactic medium ( igm ) .
the tight relation between temperature and density in this regime is maintained by the competition between adiabatic expansion cooling and photo - ionisation heating .
the plume of gas with @xmath53 and @xmath54 is comprised on the other hand of shock - heated gas in virialised haloes and , at the lower density end , in and around filaments .
the narrow , downward sloping locus of gas at @xmath55 and @xmath56 represents radiatively cooled , dense gas in galaxies .
the cooling times at these densities are short , so that gas remains close to its equilibrium temperature where photo - ionisation heating balances radiative cooling .
this temperature is a slowly decreasing function of density and lies close to @xmath57 .
finally , once the gas becomes very dense and reaches the star formation threshold , we describe the gas by an effective equation of state representing the mean thermal energy density of a two - phase medium of hot and cold gas .
this effective equation of state results in the upward sloping high density line of gas , which represents the ism in galaxies .
the pressurisation by supernova feedback in our subresolution model prevents this gas from fragmenting under self - gravity . as fig .
[ fig : rho_t ] demonstrates , arepo and gadget lead to a qualitatively very similar @xmath58 phase - space diagram .
this demonstrates that the properties of the gas distributions are overall quite comparable in both codes .
however , although there is general broad agreement between the simulations , there are also some striking differences .
first of all , the distribution of hot gas extends to higher densities in the arepo runs compared to the gadget simulations ; i.e. there is more hot gas at higher densities in the arepo simulations .
we note however that in terms of total mass this is not very significant .
more important , gadget appears to have more hot gas in general , which can be inferred from the slightly more extended yellow region in the phase diagrams .
this is consistent with the temperature structure in the surroundings of the galaxy shown in fig .
[ fig : projections ] .
the arepo runs also exhibit a more pronounced cooling feature around @xmath59 , which corresponds to a local minimum of the cooling curve between the two line peaks of hydrogen and helium .
although this feature is readily visible in the phase - space diagrams of the arepo runs , the actual gas mass populating that feature is very small .
the faint stripe - like features seen in the top - left and bottom - left panels of fig .
7 below the effective equation of state arise from numerical inaccuracies in the temperature evolution of some cells around star - formation sites . here
the temperature can sometimes scatter to very low values , which is one incarnation of the accuracy problems associated with supersonic cold fluid motions . in our multi - phase model , the temperature of these cells
is then relaxed back to the effective equation of state temperature @xmath60 on a timescale @xmath61 given by equation 12 of of @xcite .
the simulation timestep @xmath62 of these cells is small against @xmath61 , such that the temperature of the cell at the end of a relaxation step is @xmath63 . because @xmath64 ( see * ? ? ? * ) and due to the power - of - two hierarchy of possible values for @xmath62 , a set of parallel stripes of cells spaced by a factor of 2 results .
the mass in these cells is however negligibly small , so that this effect has no influence on the dynamics . to quantify the gas mass content in the different regions of the @xmath49@xmath50 phase - space diagrams in fig .
[ fig : rho_t ] , we introduce the following cuts in the @xmath49-@xmath50 plane : ( 1 ) diffuse cool gas with @xmath65 and @xmath66 , ( 2 ) diffuse warm gas with @xmath65 and @xmath67 , ( 3 ) diffuse hot gas with @xmath65 and @xmath68 , and finally , ( 4 ) dense gas with @xmath55 . in fig .
[ fig : gas_phases_mass ] , we first show the different gas mass fractions of each of these four phases .
all runs show the behaviour we anticipated earlier ; i.e. there is the general trend that the cold gas fraction decreases and the warm / hot gas fraction increases with time . the hot gas fraction ( lower left panel ) increases due to shock heating in haloes .
interestingly the details vary quite strongly between the different numerical schemes .
the gadget runs typically show more hot and warm gas , and accordingly produce less dense gas .
especially the mass fractions in the hot phase are very different between the two schemes : for most of the time the mass difference here is larger than one order of magnitude .
we will show below that these two findings are also consistent with the results for the star formation histories of the different runs .
we note that the same trends are also found for the lower resolution simulations , but the quantitative numbers differ slightly .
although not fully converged , there is more very dense gas ( lower right panel ) in arepo than in gadget at all resolutions , and also the amount of cold gas ( upper left panel ) is larger in arepo .
these results confirm the visual impression obtained from fig . [
fig : projections ] , where we also saw that the hot atmosphere is smaller in arepo , but the cold gas disk is larger .
next , we consider in fig .
[ fig : gas_phases_volume ] the corresponding volume fractions of these gas phases .
the volume fractions of the gadget runs are based on the specific volume of individual sph particles : @xmath69 .
however , sph is not volume - conserving in the sense that the sum of the specific volumes of all sph particles will not sum up to the total simulation volume .
in finite - volume methods like the one used by arepo , this does not occur ; all the cell volumes add up to the correct simulation volume .
for the comparisons of the volume fractions we therefore do not calculate volume fractions with respect to the total simulation box volume , but with respect to the sum of all individual particle / cell volumes .
this does not influence the volume fractions of arepo , but changes the sph volume fractions of gadget , which would otherwise not add up to unity .
the general trends we find for the volume fractions in fig .
[ fig : gas_phases_volume ] are similar to those for the mass fractions in fig . [
fig : gas_phases_mass ] .
however , we can now clearly see that the hot atmospheres are not only strongly reduced in mass in arepo , but also that their volume fraction is significantly smaller compared to gadget .
this is in agreement with the qualitative results shown in fig .
[ fig : projections ] and fig .
[ fig : collage_gas_z2 ] . the same is true for the volume fraction of the cold gas , which is however not as well - converged . to focus more on the density structure ,
we marginalise fig . [
fig : rho_t ] over temperature and express the densities as physical hydrogen number densities .
we hence derive the hydrogen number density probability distribution functions for the full simulation volume , which are shown in fig .
[ fig : density_pdf ] . the left panel accounts for all gas particles / cells , whereas the right panel shows only those with active star formation . in our star formation model ,
the density threshold for the onset of star formation is set to @xmath4 , and therefore the right hand panel occupies only this high density region .
interestingly , gadget shows at all resolutions a significantly reduced gas fraction slightly below the star formation density threshold .
we checked that the reduction of the gas mass in this region is caused by the more extended disks in arepo ; i.e. the additional gas found in arepo at these densities stems from disk material .
presumably , the ` gap ' phenomenon found in sph across strong contact discontinuities @xcite , such as the one we have here , also contributes to the different breakdown in that density range .
the other parts of the density distributions look rather similar for the two simulation methods , however .
we verified that the gas slightly below the threshold is predominantly responsible for the more extended disks in arepo .
this gas effectively fills the gap which is present in sph .
the galaxies in gadget typically lead to a peak in the density pdf above the star formation threshold instead of a more plateau - like feature due to the extended gas disks in arepo .
therefore , the difference in the density pdf is a reflection of significantly different galaxy radii for the different hydro - schemes ( see also paper ii ) . , @xmath44 and @xmath70 as indicated . at higher
redshift ( @xmath70 ) the two curves agree reasonably well in shape .
but at @xmath44 there is already a trend visible , because the arepo distribution is biased towards slightly younger stars .
this effect becomes even more pronounced towards lower redshift and is clearly visible in the distribution at @xmath5 . at that time
the arepo result is significantly different from gadget and shows a substantially larger fraction of young stars .
the reason for this is the larger overall star formation rate at late times , as shown in fig .
[ fig : sfr_hist ] .
the different stellar age distributions are also responsible for the bluer appearance of the stellar populations of the arepo galaxies shown in fig .
[ fig : collage_stars_z2 ] and fig .
[ fig : collage_stars_z0 ] .
[ fig : stellar_age],scaledwidth=47.5% ] an important prediction of our simulations is the global star formation history ( sfh ) , which encodes key information about the overall efficiency of the galaxy formation process .
the sfhs of our different simulation runs are presented in fig .
[ fig : sfr_hist ] .
the top left panel shows the sfh as a function of lookback time , whereas the lower left panel plots it on a logarithmic scale as a function of redshift .
the corresponding integrated star formation histories are shown in the right panels .
the most obvious trend with increasing resolution is that more high-@xmath71 star formation in small haloes is resolved @xcite .
this is simply due to the better mass and spatial resolution for smaller haloes , which are in part not even seen in the coarser simulations .
this trend with resolution also shifts the peak in the star formation rate density to higher redshift with increasing resolution .
note however that there is little cosmic time at these high redshifts , so that the mass of stars formed there is small compared to the present stellar density .
once the simulations resolve all haloes with virial temperature of @xmath72 and above ( necessitating a dark matter particle resolution of about @xmath73 ) , we expect however that the trend with resolution will stop and all ordinary star formation will be resolved ( apart from additional pop - iii star formation , which is not included in our models ) . after the high - redshift peak of star formation , the star formation rates decline and converge particularly well at lower redshifts for the sph - based gadget simulations . in this regime , most of the star formation is contributed by higher mass haloes @xcite , for which our subresolution model produces converged results already at moderate resolution
. this excellent convergence does not quite carry over to the moving mesh calculations with arepo , where all resolutions slightly vary in their late time star formation rates ( for further explanation see section 4.3 and paper iii ) . however , the mesh - based results are still very close to one other , and more importantly , they lie significantly higher than the sph simulations , which is clearly visible from the inset in the upper left panel of fig . [ fig : sfr_hist ] showing the ratio of the individual sfhs to the mean .
we emphasise once more that this systematic difference is a result only of the different hydrodynamical schemes , since both codes use the same star formation model and calculate identical star formation rates for a given gas density .
there is a minor technical difference in how stellar particles are created , but this does not affect the outcomes , as we have explicitly checked in numerous test calculations of isolated haloes and galaxies ( see paper iii ) .
the latter do show very good agreement in the star formation rates between gadget and arepo , so the difference we find in fig .
[ fig : sfr_hist ] must have a cosmological origin . to understand better whether this increased star formation rate affects all halo masses or whether it is dominated by a certain halo population , we plot in fig .
[ fig : sfr_fof ] the star formation rates as a function of fof group mass for our highest resolution simulations .
solid lines in fig .
[ fig : sfr_fof ] represent the median values and shaded regions show the @xmath74 range of the distribution .
the insets show the specific star formation rates . in both cases , and for all redshifts shown , the scatter around the median relation is rather small . at high redshift ( @xmath75 ) , the differences between the sph and moving - mesh calculations mainly occur at the intermediate and low mass end , where gadget produces more stars than arepo .
this is in agreement with the general trend found for the global sfh , where gadget shows a slightly higher star formation rate at high redshift .
we note that the higher specific star formation rates in arepo at late times are due to a significantly more efficient cooling .
we point out that one reason why gadget shows at high-@xmath71 a larger star formation rate ( sfr ) at a given resolution than arepo is related to the fact that already at this early time the disks in arepo are slightly more extended and hence have a lower gas surface density ( as demonstrated in paper ii ) .
therefore , the star formation rate is lower compared to the denser , more blob - like galaxies formed in the sph calculations .
we note that this discrepancy between gadget and arepo decreases once the resolution in sph becomes high enough , but it still induces a noticeable delay in arepo s sfr - peak compared to gadget , although the peak amplitudes agree quite well . ) of the gas for the different simulations , where @xmath76 and @xmath77 are the cooling and heating rates per unit mass .
we set the net cooling to zero for particles / cells that are subject to net heating .
the rates of all arepo runs are higher than those of the gadget simulations at late times .
this is directly related to the higher sfrs of all arepo runs compared to gadget , as seen at late times in fig .
[ fig : sfr_hist ] . at high redshift ,
the cooling rates of arepo are lower than those of gadget for the intermediate and the lowest resolution .
this is a consequence of the more compact disks in gadget at these high redshifts .
this cooling rate difference decreases once we reach the highest resolution . in that case
the cooling rates of gadget and arepo agree better at high redshifts , but still show quite large differences at low redshifts , which also produces the different sfr history .
[ fig : cooling_rates],scaledwidth=48.0% ] towards lower redshifts , a gradual trend of a different nature appears : while arepo and gadget start to agree better at low masses , the schemes deviate more and more strongly towards the high mass end . at @xmath5 ,
there is a nearly perfect overlap of the star formation rates below @xmath78 , but fof groups above this mass form significantly more stars in the arepo simulations compared to gadget runs .
interestingly , this difference in large haloes first starts to show up at around @xmath70 ( upper right panel ) , which roughly coincides with the time when the global star formation rates start to deviate , as can be seen from the lower left panel of fig .
[ fig : sfr_hist ] .
this implies that the massive end of the halo population is mainly responsible for the higher star formation rate in the moving mesh calculations at late times .
these massive systems form at late times and systematically create more stars in arepo than the corresponding gadget calculations , driving the global sfh to a different behaviour in the two codes .
the only way such a difference in the sfh can occur is a more efficient cooling behaviour in arepo in haloes above @xmath79 , allowing more gas to accumulate at the bottom of halo potentials and above the density threshold for star formation .
this interpretation also agrees with our findings for the mass fractions and volume fractions in the different gas phases , where gadget shows at all resolutions significantly more hot gas than arepo .
furthermore it agrees with the gas density pdfs in fig .
[ fig : density_pdf ] , which show that arepo has a larger fraction of dense gas slightly below the star formation density threshold .
as we will discuss in the next section , the culprit of the reduced cooling in gadget lies in different hydrodynamical dissipation rates in the outer parts of haloes compared with arepo . in paper iii of this series ,
we furthermore show that arepo s more accurate treatment of hydrodynamical fluid instabilities strongly affects the stripping of gas out of gas clumps that fall into haloes .
this in turn leads to significant differences in how the thermodynamic structure of haloes is affected by the hierarchical merging process , contributing to the cooling differences .
we note that the different star formation histories also result in a different distribution of stellar ages in individual galaxies .
this can already be seen in fig .
[ fig : collage_stars_z2 ] and fig . [ fig : collage_stars_z0 ] , where not only the morphology of the stellar distribution is different , but also the average age as can be inferred from the colour scale .
the gadget galaxies look clearly redder overall than the arepo galaxies , which are on average bluer . to quantify this in more detail ,
we show in fig .
[ fig : stellar_age ] distribution functions of the formation redshifts of the stellar populations at @xmath5 , @xmath44 and @xmath70 , for our two highest resolution simulations .
this figure demonstrates that at high redshift ( @xmath70 ) the distributions agree reasonably well in shape .
but at @xmath44 , a clear trend is already visible , because the arepo distribution becomes biased towards slightly younger stars .
this has to do with the slight shift of the peak of the sfh and the different late time behaviour of the sfh found in fig .
[ fig : sfr_hist ] .
the difference in the distributions becomes even more pronounced towards lower redshift and is clearly evident at @xmath5 . at that time
the arepo distribution is quite distinct from gadget and shows a significantly larger fraction of young stars .
the reason for this lies in the larger star formation rate at late times , as discussed above .
the differences between arepo and gadget discussed so far point towards a different cooling behaviour between the two codes . to demonstrate this more clearly , fig .
[ fig : cooling_rates ] shows the total net cooling rate ( @xmath80 ) , where @xmath76 and @xmath77 are the cooling and heating rates per unit mass as a function of redshift , accounting for all gas in our simulations . in the case of net heating
we set this rate to zero .
the figure demonstrates that at late times , the cooling rates of all arepo runs are higher than those of corresponding gadget runs .
we note that this is true for all three resolutions , yielding a clear separation between the set of blue curves ( representing the sph simulations ) and the red curves ( representing the moving mesh calculations ) .
interestingly , the difference partially reverses at high redshift , where the cooling rates of arepo at the intermediate and the low resolution are lower than those of gadget .
this is a consequence of the more compact galaxies in gadget at these high redshifts , which result in larger densities and therefore more cooling radiation .
this high-@xmath71 cooling rate difference decreases and nearly disappears once we reach the highest mass resolution . in that case
the cooling rates of gadget and arepo agree reasonably well at high redshifts , as shown in fig .
[ fig : cooling_rates ] .
in contrast , the quite large and systematic difference at low redshifts persists independent of resolution .
further insight into the origin of the cooling difference is provided by fig .
[ fig : density_vs_coolingrate ] , which shows the mean mass - weighted net cooling rates ( @xmath81 ) at redshifts @xmath44 ( left panel ) and @xmath5 ( right panel ) as a function of hydrogen number density .
if the temperature structure of the gas at fixed density would agree between all simulations , then the mean mass - weighted cooling rates should also agree for a given hydrogen number density .
but as fig .
[ fig : density_vs_coolingrate ] shows , this is not the case and some systematic differences between gadget and arepo are present .
interestingly , the curves in fig .
[ fig : density_vs_coolingrate ] differ however only over a limited density range corresponding to the diffuse gas in haloes , beginning at the star formation density threshold ( marked by the vertical green line ) and extending to lower densities . in that density range
, arepo shows a significantly larger mean cooling rate at all resolutions , which must be due to a slightly lower gas temperature in the mean at these densities .
we note that the magnitude of the difference is however redshift dependent and becomes smaller as we go to higher redshifts , as seen by comparing the @xmath5 and @xmath44 panels in fig .
[ fig : density_vs_coolingrate ] . at other hydrogen densities , including the range not shown in fig .
[ fig : density_vs_coolingrate ] , the sph and moving - mesh cooling rates agree well for all our runs . above the threshold for star formation
this is however not surprising , because here the effective equation of state produces a tight correlation between density and temperature ( as seen in fig .
[ fig : rho_t ] ) .
this leads to a nearly one - to - one mapping of density to temperature in this regime , and explains the two - peak structure of the curves in fig .
[ fig : density_vs_coolingrate ] at high densities , which simply reflects the primordial cooling curve with its characteristic hydrogen and helium peaks . further evidence for a different temperature structure in the two simulation runs is provided by fig .
[ fig : mean_temp ] , which plots the mean mass - weighted temperature in the whole simulation volume as a function of time . in the inset
we divide each curve by the mean of all the curves to emphasise deviations .
we find that the mean temperature of gadget simulations is @xmath82 higher than that of the arepo runs .
the higher temperature can explain the reduction in cooling emission of the sph simulations relative to our moving - mesh calculations , and it is consistent with our above findings for the sfr evolution and the cooling emission .
it thus appears that the origin of the low redshift discrepancy must lie in a different efficiency of non - adiabatic heating processes in the gas , as this is required to explain differences in the temperature distribution at a given density .
+ + it is widely appreciated that the intracluster medium of galaxy clusters is partially supported by subsonic turbulence @xcite , which is created by curved accretion shocks around the haloes , and more importantly , by the hierarchical infall of structures .
indeed , a number of numerical studies have analysed turbulence in the intracluster medium @xcite
. a similar level of turbulence can also be expected in smaller haloes that are large enough to support quasi - hydrostatic atmospheres .
it has further been shown that shock heating is not only important in strong shocks at the outer accretion radius , but is also significant in weaker flow shocks that occur in large parts of the halo volume @xcite .
physically , the dissipation associated with shocks and with the decay of turbulence occurs through microphysics on very small scales . in our numerical approach ,
we neglect the physical viscosity ( which is assumed to be effectively zero on all resolved scales ; this is why we employ the euler and not the navier - stokes equations ; see e.g. @xcite for a navier - stokes implementation of arepo ) and account for the necessary dissipation through numerical viscosity . in sph
this is provided explicitly in terms of an artificial viscosity , whereas in arepo it is introduced implicitly through the solutions of the riemann solver and cell - averaging .
the good conservation properties of sph allow it to capture one - dimensional shock waves quite well , even though the post - shock velocity field typically shows substantial noise .
this already hints that sph may not be particularly accurate for subsonic flow phenomena , such as the turbulence we expect in the virialised gas of newly formed haloes .
indeed , @xcite have systematically compared driven isothermal subsonic turbulence for gadget and arepo , using the same versions of the codes we employ here .
they find that sph fails quite badly to account for a turbulent cascade in the subsonic case , whereas a kolmogorov scaling is obtained for the moving - mesh code .
this happens despite identical driving fields in both cases , and is ultimately caused by inaccurate gradient estimates in sph and different dissipation as a function of scale .
the sph simulations dissipate most of the energy already close to the driving scale in the subsonic case , whereas in arepo efficient dissipation happens only on much smaller scales , so that a self - similar turbulent cascade can develop over some inertial range .
in addition , @xcite find that sph exhibits a second maximum in the dissipation on very small scales of order the mean interparticle separation . here the subsonic noise of sph is dissipated by the artificial viscosity .
interestingly , the total amount of dissipation on these scales is quite independent of the artificial viscosity value itself .
while a higher viscosity reduces the amplitude of the small - scale sph noise , the dissipation rate on these scales still remains roughly the same , suggesting that the cause of the small - scale noise , which originates in errors in the pressure gradient estimates , can not be reduced effectively with a different viscosity setting .
there are good reasons to expect that these differences in the dissipation properties of the hydrodynamical schemes may also induce important effects for the thermodynamic structure of cosmological haloes , which in turn can easily give rise to variations of the amount of gas that cools out of these haloes . to examine this further ,
we have carried out two auxiliary non - radiative simulations at resolutions of @xmath83 and @xmath84 particles / cells .
these simulations use the same box - size and parameters as our galaxy formation runs , except that cooling and star formation were not included .
the use of non - radiative simulations allows us to cleanly employ the method of @xcite for measuring the instantaneous dissipation rate of particles and cells , respectively . in the case of sph , this can simply be done by measuring the work per unit time done by the artificial viscosity forces , which represents the sole source of entropy generation in this method . for the moving - mesh code
, we instead also advect the thermodynamic entropy between the cells , and then measure the rate of entropy generation by comparing the state of a cell at the end of a timestep computed adopting energy conservation with the state expected if the entropy would have remained constant .
the inferred rate of entropy production can then also be converted into a fiducial heating rate .
haloes contained in different mass ranges , from more massive ( top ) to less massive ( bottom ) , as labelled .
note that the bottom panel is identical to the last panel in fig .
[ fig : dissipationrate1 ] , which is here repeated for ease of comparison .
the solid lines show results for the @xmath84 non - radiative simulations , while the dashed lines give the equivalent @xmath83 results .
[ fig : dissipationrate2],title="fig:",scaledwidth=39.0% ] + haloes contained in different mass ranges , from more massive ( top ) to less massive ( bottom ) , as labelled . note that the bottom panel is identical to the last panel in fig .
[ fig : dissipationrate1 ] , which is here repeated for ease of comparison .
the solid lines show results for the @xmath84 non - radiative simulations , while the dashed lines give the equivalent @xmath83 results .
[ fig : dissipationrate2],title="fig:",scaledwidth=39.0% ] + haloes contained in different mass ranges , from more massive ( top ) to less massive ( bottom ) , as labelled . note that the bottom panel is identical to the last panel in fig .
[ fig : dissipationrate1 ] , which is here repeated for ease of comparison .
the solid lines show results for the @xmath84 non - radiative simulations , while the dashed lines give the equivalent @xmath83 results .
[ fig : dissipationrate2],title="fig:",scaledwidth=39.0% ] + in fig .
[ fig : dissipationrate1 ] , we compare stacked profiles of the spherically averaged dissipative heating rate of haloes obtained in this way . for the plots , we have selected all haloes in the virial mass range @xmath85 , where @xmath86 refers to the mass inside a radius @xmath87 that encloses 200 times the critical density .
we have then placed 20 logarithmic bins onto the radial range @xmath88 to @xmath89 , and produced mass - weighted averages of the dissipation rate per unit mass , @xmath90 , that we measured for each particle / cell . in order to allow a calculation of an average profile by stacking the haloes
, we express the dissipation rate in dimensionless units by multiplying it with the hubble time at the given redshift , and by normalising it to @xmath91 of the particular halo , where @xmath92 is the circular velocity at the virial radius . in the individual panels of fig .
[ fig : dissipationrate1 ] , we show results for different redshifts , ranging from @xmath75 to @xmath5 , and we compare arepo with gadget at the two resolutions considered here . we see that there is a clear systematic difference in the dissipative heating rates as a function of halo - centric distance .
arepo produces more heating in the infall region directly outside the virial radius whereas gadget shows more heating in the outer parts of virialised haloes , where most of the gas mass is located .
this systematic difference is present at all redshifts in large haloes .
we also note that the nature of the difference is robustly preserved at different resolutions , even though there seem to be small residual trends with spatial resolution .
interestingly , at fixed halo mass , the relative level of dissipation in the innermost parts of haloes increases in the mesh - code with time , which may be related to the build - up of a turbulent entropy core in these haloes . in fig .
[ fig : dissipationrate2 ] , we consider dissipation profiles as a function of halo mass at @xmath5 , which however does not reveal any clear trend with halo mass .
our interpretation of the systematic difference revealed by figs .
[ fig : dissipationrate1 ] and [ fig : dissipationrate2 ] is that : ( 1 ) shocks are captured more efficiently by arepo in the accretion regime of haloes , and ( 2 ) arepo strips gas out of infalling objects more efficiently there ; both effects contribute to a stronger dissipation in this region .
in contrast , sph stops infalling gas less efficiently , leading to more dissipation at smaller radii . a further contributor to
the heating there lies in the dissipation of the subsonic noise in sph , which is constantly recreated in this region by tapping free energy from the disturbances that impinge on the haloes from the outside . as a net result of this dissipation of subsonic noise in the outer parts of haloes , the sph simulations end up hotter
overall ( see also fig .
[ fig : mean_temp ] ) .
furthermore , this difference in the heating occurs in a region where one expects the cooling radius of many haloes in cosmological radiative simulations , thereby directly affecting the amount of cold gas produced in the quasi - hydrostatic cooling flows in large haloes .
we note that the measurement of the dissipation rate is difficult due to its high time variability .
we use above stacked profiles of small sets of halos to mitigate this problem , but even then the inner logarithmic bins average over much smaller gas mass than the outer parts .
this implies that the profiles in the inner @xmath93 of the virial radius are relatively noisy .
but this inner part is largely irrelevant as the impact on the cooling rates comes from larger radii as we discussed above .
another reason why differences in the global star formation rate occur between arepo and gadget lies in the different accuracy with which hydrodynamical fluid instabilities are treated . in particular
, we expect that for the moving - mesh code cold gas can be stripped more efficiently from galaxies as they fall into larger haloes , and this gas is then mixed with the halo s diffuse gas , which also lowers its temperature . to demonstrate this difference
more explicitly in our cosmological runs , we performed special test simulations of a @xmath94 box with @xmath95 sph particles / cells . in fig .
[ fig : gas_balance ] , we show the sum of the total amount of star - forming gas mass and stellar mass in this box as a function of redshift .
we note that an sph particle / cell is star - forming if its density is higher than the threshold of the subresolution star formation model .
the solid lines show the result of calculations with our standard star formation prescription , i.e. these simulations are equivalent to the main runs presented in the paper . however , for the dashed lines we turned off the creation of stellar particles , keeping everything else the same . in this case , the cold gas accumulates in the galaxies and is supported by the effective equation of state expected for gas at these densities , except that the gas is not depleted and does not turn into stars at the normal rate . for these runs ,
the mass in the figure therefore refers to the total amount of star - forming gas in these simulations ( no stars are present ) .
interestingly , the solid and dashed curves overlap well for gadget ( blue ) in fig . [
fig : gas_balance ] , showing that once gas has overcome the density threshold it will never return to lower density . stripping of gas out of galaxies does not appear to happen in any significant way , otherwise we would expect that the run without star formation should end up with a smaller amount of collapsed baryons .
instead , the sph result is consistent with no stripping at all , i.e. once a gas particle has cooled , it never returns to lower density even though the galaxy may be subject to substantial shear flows upon halo infall .
this behaviour has also been found by he & springel ( 2011 , submitted ) in a comparison of galaxy - wind interactions in sph and the vph technique @xcite .
for arepo the situation is very different . as fig .
[ fig : gas_balance ] shows , the sum of stellar mass and star - forming gas mass is larger than the amount of star - forming gas in the run without star particle creation . here
the relevant fluid instabilities for stripping ( like the kelvin - helmholtz instability ) are significantly better resolved and facilitate a substantial mass loss of infalling overdense blobs / subhaloes . because of this , the sum of the mass of stars and of star - forming gas is higher in arepo for the run with star particle creation than for the fiducial run where this is suppressed . in the latter simulation ,
there is simply more gas around that can be stripped again , while in the run with star particle creation , baryons that have been converted to stars do of course not suffer from fluid instabilities anymore .
apart from affecting the cooling in haloes , this difference in the stripping and mixing efficiency also modifies the dynamical friction timescale of infalling gas clumps , as we examine in more detail in paper iii of this series .
various modifications have been proposed to the conventional implementation of sph as in gadget to improve its reliability in certain cases .
for example , changes to the computation of smoothed pressure gradients @xcite or the addition of an artificial thermal conductivity to the equations of motion ( e.g. * ? ? ?
* ) enable sph to better handle shearing interfaces and the onset of kelvin - helmholtz instabilities . also , a number of studies focused on the development of improved artificial viscosity parameterisations in sph codes @xcite , with the goal of reducing the numerical viscosity away from shocks .
given the results presented in this paper , it is tempting to argue that one or several of these modifications of `` standard sph '' may lead to much better agreement with the moving - mesh results or even resolve the discrepancies . while we can not exclude this possibility , we note the above refinements are ad hoc in the sense that they are designed to mitigate against particular problems in some circumstances , and may have unwanted side - effects in other situations . while these side - effects may often be benign ,
the proposed modifications of sph do not address other , more generic problems with this technique . in this section , we summarise these issues in order to put our findings into context with the recent literature on sph techniques .
one issue that has rarely been discussed in the cosmological community concerns the convergence properties of sph .
this is manifested in at least two ways , the first having to do with local smoothed estimates in sph , and the second is related to the fact that sph is not formally lagrangian .
below , we first address the question of under what conditions convergence is expected in sph and how this influences the relation between spatial resolution and the total particle number .
then , we discuss the pseudo - lagrangian character of sph and its implications for defining convergence with this method .
finally , we briefly suggest ways in which these issues might be dealt with . incidentally
, they all would have the effect of making sph resemble a moving - mesh scheme like arepo .
box with @xmath95 sph particles / cells .
the @xmath96-axis shows the total amount of star - forming gas and stellar mass in the simulation volume .
solid lines show the result of a calculation with our standard star formation prescription . for the dashed lines we turned off the creation of stellar particles ;
i.e. there are no stars formed but the gas still cools and is supported by the multiphase equation of state .
the solid and dashed curves overlap well for gadget , which shows that once gas overcomes the density threshold it either forms stars or stays above the threshold .
there is no loss of star - forming gas mass . for arepo , the amount of stellar mass and star - forming gas mass is larger than the amount of star - forming gas if stellar particle creation is turned off
this is because arepo resolves fluid instabilities better and therefore allows significant gas mass loss from infalling satellites . here
gas that once has been above the threshold for star formation may well be mixed with halo gas and reach much lower densities again . on the other hand ,
once stellar particles are created from dense gas , the total amount of baryonic material that can be lost from subhaloes is reduced , which explains the difference seen in the two moving - mesh calculations.,scaledwidth=47.5% ] for a real fluid , we define the density , @xmath97 , to be a continuous function of space , determined by averaging over many molecules locally around @xmath98 . in order to formulate a discrete sph counterpart to this ( and this applies to all fluid variables , as well as the equations of motion ) a two - step procedure is applied . first , we introduce a smoothed version of the original density field , @xmath99 , by convolving @xmath97 with a smoothing kernel @xmath100 according to @xmath101 where the integral is over all space and @xmath102 is the smoothing length .
second , the continuous smoothed field @xmath103 is replaced by a discrete quantity , @xmath104 , so that it can be represented computationally .
this is done by regarding @xmath105 as a mass element @xmath106 and partitioning the fluid into a set of @xmath107 discrete mass elements so that the above integral can be approximated by a discrete sum : @xmath108 where @xmath109 is the mass of fluid element @xmath110
, @xmath111 is its spatial location , and the expression allows different fluid locations / elements to have different smoothing lengths . in principle , the sum extends over all @xmath107 fluid elements .
however , in order that the smoothed density @xmath112 approaches the continuum limit represented by @xmath97 , it is necessary that @xmath100 is spatially localised .
so in practice a localised kernel with compact support is adopted , implying that the discrete sum extends not over all @xmath107 fluid elements but over a subset @xmath113 of those neighbouring a certain point in space .
now , we can ask about the conditions that must be satisfied for the smoothed , discrete version of the density field to asymptote to the actual continuum limit .
that is , under what circumstances is the following true ?
@xmath114 consider the second of these requirements first . in order for @xmath115
we must have @xmath116 as @xmath117 , as implied by the definition of @xmath112 in the above integral .
this also requires that @xmath118 , otherwise there exist not enough particles around a given spatial location to form smoothed averages .
turning to the first requirement above , @xmath119 will be satisfied only if at the same time we enforce the condition @xmath120 .
this has not been emphasised in the cosmological literature , but from the defining relation above for @xmath121 it is clear that the discrete approximation will not asymptote to the smoothed density field unless this limit is taken . or , to put it another way ,
if the number of neighbours is held fixed as @xmath122 , there will be a constant source of error present on the most well - resolved scales that will not vanish as the resolution and number of particles is increased indefinitely ( see also * ? ? ?
* ; * ? ? ?
( for a discussion of this requirement in the context of elasticity , see e.g. @xcite . )
to the best of our knowledge , the optimal way to enforce this requirement has not been established . however , there is little doubt that this must be addressed in order that sph provides properly convergent solutions , as argued recently by @xcite : `` when performing a convergence study using sph , it is important to vary both the number of particles ... as well as the ratio of smoothing length to particle spacing [ to increase the number of neighbours ] .
[ otherwise ] the error due to the summation interpolant [ remains ] constant [ as the total particle number is increased ] . ''
we emphasise that this requirement is not fundamentally an issue with sph , but simply reflects the fact that a numerical approximation to an integral in the form of a discrete sum will not converge to the true answer unless the number of points where the integrand is sampled is made larger .
an increase in the number of neighbours is also warranted to reduce the gradient errors in sph , which give in fact rise to a sizable `` zero - th order error '' @xcite .
these gradient errors seriously degrade the accuracy of sph in subsonic flow problems and are a primary cause for problems in the representation of subsonic turbulence @xcite . however , in practice , simply increasing the number of neighbours is met by a serious obstacle , because it invokes the so - called clumping instability in which particles located in the inner parts of the kernel are pushed together by the pressure forces of the surrounding particles , forming little clumps that frustrate attempts to reduce the error of the kernel sums in this way . partially circumventing this problem and allowing a higher neighbour number requires different kernel shapes than are commonly employed @xcite , but such kernels merely have different stability bands that still impose severe restrictions on the number of neighbours that may be used .
in essence , due to the clumping instability , establishing formally convergent sph solutions is an unsolved problem because the path of increasing the neighbour number is blocked in practice .
the requirement that the number of neighbours should increase to reduce errors as the total number of particles is made larger has also serious implications for the efficiency of sph . as an example , consider a uniform fluid in a cube of side - length @xmath123 represented by @xmath107 sph particles .
( for more general circumstances the argument generalises if we think about small scales on which the fluid properties are nearly uniform . )
the mean separation between the particles , @xmath124 , is @xmath125 if we take @xmath126 , as has been done conventionally in sph codes used in cosmology , then the smoothing length will shrink in proportion to @xmath124 and , so , the number of neighbours will be nearly constant . in a sense , this maximises the spatial resolution , but leaves a fixed source of error in the discrete sums @xcite , so this path is in principle not appropriate for achieving proper convergence with sph .
if , instead , we set @xmath127 then the fundamental relation between @xmath107 and @xmath102 becomes @xmath128 .
what are the conditions on @xmath129 so that we can meaningfully construct a procedure that converges numerically ? clearly , @xmath130
, otherwise @xmath102 will not decrease more slowly than @xmath124 as @xmath107 increases . also , we must have @xmath131 , as the choice @xmath132 would mean that the spatial resolution is constant , independent of @xmath107 .
the optimal value for @xmath129 is undetermined .
one physically appealing choice would be @xmath133 , because then @xmath102 would be the geometric mean between the mean interparticle separation , @xmath124 , and the size of the system , @xmath123 . in that case
, the relation between @xmath107 and @xmath102 becomes @xmath134 , implying a steeply rising computational cost with resolution .
it is possible that a slightly larger value of @xmath129 could be preferred in terms of cpu costs , but it is clear , however , that in order to extend the dynamic range in spatial scales resolved by sph in a way that guarantees eventual convergence , it is absolutely necessary to increase the number of particles more aggressively than has been advocated previously .
the dilemma , of course , is that the clumping instability makes this highly problematic . ignoring the clumping issue for the moment , the implications for the convergence rate of sph are in any case important . when the common practice in the field is followed and the number of neighbours is held constant , reported a global l1 error for a vortex flow ( the gresho test ) that scales as @xmath135 with resolution , as opposed to @xmath136 for arepo .
we can then ask how the ratio of the cpu - time cost of simulations with the two schemes varies with the size of the error . for simulations in 3d , the computational cost scales for both codes as @xmath137 , where three powers of @xmath102 come from the spatial dimensions , and one additional power enters due to the reduction of timesteps for better resolution .
this then implies that the cost ratio scales as @xmath138 with the error of the calculation .
reducing the error by a factor of 10 requires therefore about a @xmath139 times higher effort in sph than in arepo .
this conclusion is problem dependent and can also be affected by specific details of the sph algorithm , like the kernel shape . however , the question of computational efficiency of different numerical schemes is best phrased in terms of such convergence rate comparisons . here
the monte carlo character in the approximation of the sph kernel sum invariably introduces a serious disadvantage for sph compared with the moving mesh approach , as outlined above .
a related conclusion , but focusing on the efficiency impact of the artificial viscosity parameterisation in sph , was reached by @xcite .
if the goal is a representation of subsonic turbulence with a certain reynolds number @xmath140 , they showed that the computational cost of sph scales at least as @xmath141 .
in contrast , in the mesh - based treatment of arepo , the dissipation scale is tied to the mesh resolution , implying a computational cost for reaching a certain reynolds number that scales as @xmath142 .
given that our cosmological simulations with arepo are only about @xmath143 slower than gadget for simulations involving self - gravity , for the same number of resolution elements , this means that at a given computational cost , arepo resolves much larger reynolds numbers than sph .
in fact , to reach the same reynolds numbers as in our present moving - mesh simulations , it appears plausible that a factor @xmath144 more sph particles than arepo cells would be required . .
the middle panels show the true gas distribution of the disk a short amount of time later . for solid body rotation
, the initial blue region retains its shape .
this is different for the differentially rotating disk , where the blue gas patch experiences the shear in the disk and gets distorted .
the panels on the right - hand side show the corresponding sph representations of the disk at this later time .
clearly , the distortion in the differentially rotating case is not captured correctly by sph since the smoothing region does not change in shape .
sph is therefore only pseudo - lagrangian , as opposed to arepo , which implements a quasi - lagrangian scheme that allows for mass exchange between cells in a manner consistent with the hydrodynamic equations of motion .
the dashed circle in the bottom middle panel illustrates a re - partitioning of the gas into sph particles at the updated time and demonstrates why mixing is suppressed in sph at the resolution scale ( see text).,title="fig:",scaledwidth=47.5% ] 0.25 cm .
the middle panels show the true gas distribution of the disk a short amount of time later . for solid body rotation
, the initial blue region retains its shape .
this is different for the differentially rotating disk , where the blue gas patch experiences the shear in the disk and gets distorted .
the panels on the right - hand side show the corresponding sph representations of the disk at this later time .
clearly , the distortion in the differentially rotating case is not captured correctly by sph since the smoothing region does not change in shape .
sph is therefore only pseudo - lagrangian , as opposed to arepo , which implements a quasi - lagrangian scheme that allows for mass exchange between cells in a manner consistent with the hydrodynamic equations of motion .
the dashed circle in the bottom middle panel illustrates a re - partitioning of the gas into sph particles at the updated time and demonstrates why mixing is suppressed in sph at the resolution scale ( see text).,title="fig:",scaledwidth=47.5% ] sph is often referred to as a lagrangian algorithm , but this is not formally correct and we suggest it is perhaps better to characterise it as a `` pseudo - lagrangian '' technique . to appreciate the reasons for this ,
consider the following example , which also highlights that this distinction is intimately related to issues of ( suppressed ) mixing in sph .
suppose gas is revolving in a thin disk on circular orbits , and imagine that the gas is partitioned at some instant into regions that are then represented by sph particles , with the smoothing done over a circular area .
each sph particle will comprise gas from an area around its centre , extending outwards in radius of order the local smoothing length , as shown for one such sph particle in the left panels of fig .
[ fig : sph_scheme ] at time zero .
consider now how the system will look like a short time later , @xmath145 , after the disk has rotated by a small angle .
if the disk is in solid body rotation , as in the upper panels of fig .
[ fig : sph_scheme ] , the fluid initially contributing to the sph particle will occupy an area identical to the smoothing region of the sph particle at time @xmath145 , as indicated by comparing the `` true '' situation in the upper middle panel with the sph version in the upper right panel . in this case , the sph representation is faithful to the actual equations of motion because the fluid contributing to the sph particle at time zero is the same as that at @xmath62 , and the region bounded initially by the sph smoothing volume has not changed its shape . however , suppose instead that the disk is rotating differentially , turning around more rapidly in the inner parts than in the outer ones ( lower panels of fig .
[ fig : sph_scheme ] ) .
the `` true '' situation shown in the middle panel tells us that the gas contributing to the area bounded by the sph smoothing volume at time zero will be stretched out owing to shear .
however , in the sph formulation , shown in the lower right panel , the material initially in the sph smoothing volume is forced to remain tied to this area and is not allowed to shear , inconsistent with the equations of motion of the real fluid .
a true lagrangian picture would allow fluid elements to be deformed in the presence of shear , but this is inhibited by the sph approach on scales smaller than those set by the smoothing procedure .
hence , while sph retains some characteristics of a lagrangian method , it does not evolve the fluid entirely in a manner consistent with the equations of motion and should , in this sense , be termed `` pseudo - lagrangian . ''
unfortunately , the error incurred by the approximations inherent to the sph formalism is highly problem dependent and can not be estimated based solely on the discretisation procedure .
for example , there is no error made in the case of a disk in solid body rotation , or for a uniform disk rotating differentially , or for gas motions that are limited to expansion or contraction in three - dimensions .
however , this is not true for shearing flows involving fluids with distinct internal properties ; here the implicit `` remapping '' involved in enforcing a spherical shape for the sph smoothing volume , which would change the local composition of the internal properties , is ignored .
this limitation is ultimately the reason why sph does not handle mixing accurately , as we illustrate in the bottom , middle panel of fig .
[ fig : sph_scheme ] .
suppose at the updated time we were to re - partition the gas into new sph particles , as indicated by the dashed circle in this frame .
the gas associated with this particle should include gas from the original particle , but also gas from the surrounding flow .
this would allow mixing to occur between the gas in the original particle and nearby parts of the flow , but is not allowed to occur in sph , by construction .
we note that arepo does not suffer from this defect . in this code
, cells are not allowed to become arbitrarily distorted , in the interests of efficiency and accuracy , and so arepo is also not formally lagrangian .
however , it still evolves fluids correctly when deviations from strict lagrangian behaviour occur by allowing for mass exchange between cells , in a manner consistent with the equations of motion .
therefore , this method should be termed `` quasi - lagrangian . ''
a related issue arises in the context of n - body simulations of collisionless systems .
there , a six - dimensional phase fluid is partitioned into particles of fixed size that move through space in a manner determined by the equations of motion . because the phase fluid initially associated with a particular particle is always tied to that particle
, the small - scale dynamics of the system will not be represented properly .
however , in this case , unlike for the example discussed above , forces are strictly long - range and are less prone to inaccuracies in the small - scale distribution of the material than pressure gradients or viscosity . in that sense ,
the local distortions in the phase space fluid have fewer dynamical consequences than for a hydrodynamical fluid in three dimensions .
nevertheless , if an accurate description of the small - scale structure of collisionless systems is essential , these effects must be accounted for at some level ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we should emphasise that whether the limitations inherent to sph lead to significant inaccuracies in the solution depend in detail on the circumstances .
sph can be expected to provide reliable results for flows that are kinetically dominated , or if the motions are controlled mainly by long - range forces . in these situations , errors in the local quantities
are sub - dominant .
for example , @xcite show that while sph does not accurately describe subsonic turbulence , in the supersonic regime , when the flow energy is mainly kinetic , gadget and arepo give similar answers . also , comparisons between sph and amr codes have yielded compatible results for the structure of the intergalactic medium , as reflected in the properties of the lyman - alpha forest ( e.g. * ? ? ? * ) . at these low densities
, gravity dominates over internal energy and , moreover , the fluid is not subject to strong shear , so the above considerations are not critical . even in some cases where shear is present
, the approximations underlying sph will not be significant , provided that gravitational forces are more important than local ones and the relevant evolution in flow properties occurs rapidly .
hayward et al .
( 2012 , in preparation ) have demonstrate this explicitly by simulating galaxy mergers using both gadget and arepo .
they find that when the gas is treated using an effective equation of state , star formation rates during the course of a merger are nearly identical between these two codes .
thus , conclusions drawn from studies of galaxy collisions with sph , such as the stellar profiles of merger remnants and their evolution ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or the survivability of disks during mergers @xcite , are likely robust to variations in the hydro solver .
unfortunately , this good level of agreement does not extend to the more complex flows associated with halos in cosmological environments , as we highlight in this paper .
there , various phases of gas will be present in close proximity , often shearing relative to one another , leading to errors in simulations done with sph like those we have identified here .
also , in the hydrostatic atmospheres of halos the hydrodynamic forces are comparable to gravity forces , and the subsonic turbulence present in these regions is affected by gradient errors in sph . without a formal convergence criterion ,
the consequences of these errors are difficult to assess , because even if a solution may plateau as the number of sph particles is increased , this by no means guarantees that the correct answer is being obtained . in particular , the solution may exhibit `` false convergence , '' appearing to reach a converged solution , while fixed sources of error from e.g. summation interpolant are still present @xcite . a case in point
is the long - standing discrepancy identified for the central cluster entropy predicted by non - radiative sph and mesh - based simulations , first identified in the santa barbara cluster comparison project @xcite .
the above considerations motivate thinking about ways to cure the defects inherent to sph so that it can provide solutions of comparable accuracy and resolution to those obtained with a moving mesh approach like arepo .
for example , the artificial viscosity typically used in sph codes to handle shocks could , in principle , be eliminated by incorporating a riemann solver into sph .
pioneering efforts along these lines have been made by @xcite and @xcite .
however , these implementations solve the riemann problem for each particle - neighbour pair , and this approach will become prohibitively expensive as long as the smoothing procedure requires averaging over an increasingly large number of neighbours to achieve convergence as the total particle number is increased .
we note that alternatively a grid could be introduced into sph just for the purposes of solving the riemann problem around each particle .
this grid would necessarily have to be adaptive , and hence resemble the voronoi tessellation used in arepo . in order to eliminate the pseudo - lagrangian character of sph and
allow for fluid elements to become distorted on scales small compared to a few smoothing lengths , as should occur in e.g. shearing flows , it is necessary to allow for mass exchange between particles in a manner consistent with the equations of motion .
this would have the added benefit of eliminating the mixing problem in sph .
first suggestions in this direction have recently been made @xcite , but the best way to formulate such mixing terms is not clear .
one possibility to unambiguously calculate the required mass flux between particles would be to utilise some kind of grid .
again , this grid would need to be adaptive , like the unstructured mesh in arepo , reinforcing the notion that our new moving - mesh approach quite naturally addresses several of the generic issues with sph .
in this study , we have presented the first cosmological hydrodynamical simulations of structure formation carried out with the novel arepo moving - mesh code . compared to the widely employed sph technique
, this method promises an important gain in accuracy in the numerical treatment of gas dynamics for similar computational costs . in order to understand the implications of these differences for galaxy formation simulations , we have evolved the same initial conditions both with arepo and the well - established sph code gadget .
both codes share the same high - resolution gravity solver , and incorporate an identical radiative cooling and star formation model .
differences in the outcome are hence tied to systematics of the hydrodynamical solvers that are used .
it is the primary goal of our paper to identify the main differences and their magnitude , as well as providing evidence for the primary sources of potential discrepancies .
galaxies at @xmath5 in the halo mass range @xmath146 to @xmath147 . clearly , the gas disks forming in the arepo simulations are typically significantly larger than those in the corresponding sph runs . , scaledwidth=47.5% ] to achieve this goal , we have considered a primary set of simulations in boxes of @xmath32 that include a basic treatment of galaxy formation physics , at various resolutions ranging from @xmath83 to @xmath148 particles / cells . in addition , we have carried out a few auxiliary simulations either in smaller boxes , or of non - radiative type , in order to more clearly measure particular numerical effects . for reasons discussed in section 2.3
, we have chosen to perform the comparisons between the codes at the same initial mass resolution , because then the computational costs are similar .
our principle findings may be summarised as follows : * the overall distribution of gas in density and temperature is broadly in agreement between sph and arepo , but there are some subtle and important differences .
in particular , the mean mass - weighted temperature in arepo is slightly smaller than in gadget .
also , the hot gas extends to somewhat lower density in gadget and hence occupies a larger volume fraction . * the cosmic star formation rate density peaks at similar values , but at slightly lower redshifts in arepo . at high redshift and at high resolution , the sfrs between the two codes are in good agreement , but towards lower redshift there are significantly stronger cooling flows in arepo , causing a correspondingly larger star formation rate .
this difference occurs primarily in haloes with mass larger than @xmath149 .
* we find that the two codes differ significantly in the dissipative heating rates within haloes , with arepo producing more dissipation in the halo infall regions , whereas sph produces higher dissipative heating throughout most of the outer regions of virialised haloes .
this difference is mainly responsible for the higher temperatures found in the sph simulations , and the correspondingly weaker cooling .
we argue that this dissipation in sph is likely to be of spurious nature , and is a combination of viscous damping of sph s inherent noise and the unphysical damping of subsonic turbulence injected into haloes in the infall regions . *
visual comparison of simulated galaxies shows considerably larger stellar disks and more extended and disky stellar distributions in arepo compared with gadget . also , the halo gas surrounding the galaxies looks smoother and less clumpy in arepo .
[ fig : scale_radii ] shows gas surface density disk scale length radii obtained from exponential surface density fits of more than @xmath150 galaxies at @xmath5 in the halo mass range @xmath146 to @xmath151 .
this clearly demonstrates that the gas disks forming in the arepo simulations are typically significantly larger than those in the corresponding sph runs .
we further note that the gas surface density of disk galaxies in the arepo simulations follow very closely exponential profiles , which is typically not the case for the sph runs .
a more detailed analysis is presented in @xcite .
we note that the large disks in the arepo simulations result even without strong feedback in the form of galactic winds or modifications of the star formation prescription .
this is also demonstrated in simplified test simulations in paper iii .
the larger extent of disk galaxies is also demonstrated in paper ii , where we do not assume exponential profiles , but rather use a cut in surface density to determine a characteristic size .
we also independently calculated half - mass radii , which show the same trends .
all these results confirm the findings reported in fig .
[ fig : scale_radii ] , i.e. that arepo produces in general larger disks than gagdet .
the above findings clearly suggest that there are significant quantitative differences caused in galaxy formation simulations by the choice of hydrodynamical technique .
it appears that the limited accuracy of sph for subsonic flow phenomena such as kelvin - helmholtz instabilities or subsonic turbulence induces important differences in the predicted properties of galaxies at low redshift , affecting both their morphology and stellar mass . because of the accuracy problems of sph for certain subsonic test problems it appears that sph can not be expected to reach highly accurate results in all regimes relevant for cosmic structure formation . at the same time
, arepo performs significantly better on these same problems , and yields generally a smaller error norm and higher convergence rate when scrutinised against problems with known analytic solutions .
we are hence confident that the arepo results presented in this paper entail a more faithful treatment of cosmological hydrodynamics , especially in view of the discussion in section 5.3 .
it is also reassuring that the arepo runs improve the predicted morphologies of simulated galaxies , as our preliminary analysis suggests . in paper
ii of this series we will back up this finding with a detailed analysis of the galaxy properties .
finally , in paper iii we provide a more detailed analysis of the relevant numerical effects responsible for the differences in mixing and cooling .
the body of these results makes it clear that arepo simulations provide a significant opportunity for the development of next generation models of galaxy formation that promise to achieve a so - far unknown combination of accuracy , dynamic range , and faithfulness to the relevant physics .
the simulations in this paper were run on the odyssey cluster supported by the fas science division research computing group at harvard university .
scaling tests were done on the ranger cluster at the texas advanced computing center ( tacc ) .
we thank paul torrey for allowing us to show fig . [
fig : scale_radii ] and elena donghia , patrik jonsson , andrew macfadyen , diego j. munoz for useful discussions .
we also thank the anonymous referee for helpful comments .
ds acknowledges nasa hubble fellowship through grant hst - hf-51282.01-a .
dk acknowledges support from nasa through hubble fellowship grant hsthf-51276.01-a .
t. , 2011 , , 413 , 271 t. , hahn o. , kaehler r. , 2011 , arxiv e - prints o. , moore b. , stadel j. , potter d. , miniati f. , read j. , mayer l. , gawryszczak a. , kravtsov a. , nordlund . , pearce f. , quilis v. , rudd d. , springel v. , stone j. , tasker e. , teyssier r. , wadsley j. , walder r. , 2007 , , 380 , 963 o. , teyssier r. , moore b. , 2011 , , 410 , 1391 d. s. , 1995 , journal of computational physics , 121 , 357 j. , hut p. , 1986
, , 324 , 446 j. e. , hernquist l. , 1992 , , 30 , 705 j. e. , hernquist l. , 1996 , , 471 , 115 j. e. , hernquist l. e. , 1991 , , 370 , l65
m. r. , burkert a. , 1997 , , 288 , 1060 a. , springel v. , 2011 , arxiv e - prints c. m. , 2006 , reports on progress in physics , 69 , 3101 t. , krongauz y. , dolbow j. , gerlach c. , 1998 , international journal for numerical methods in engineering , 43 , 785 a. j. , 2010a , arxiv e - prints , 1008.1786 a. j. , 2010b , , 495 , 33 a. j. , cole s. , frenk c. s. , baugh c. m. , lacey c. g. , 2000 , , 311 , 793 m. j. , colella p. , 1989
, journal of computational physics , 82 , 64 g. , hooper d. , silk j. , 2005 , , 405 , 279 g. r. , faber s. m. , primack j. r. , rees m. j. , 1984 , , 311 , 517 j. r. , cole s. , efstathiou g. , kaiser n. , 1991 , , 379 , 440 m. , springel v. , white s. d. m. , jenkins a. , 2010 , , 406 , 896 g. l. , norman m. l. , 1998 , , 495 , 80 s. , porciani c. , 2011 , , 411 , 1678 d. c. , xu h. , norman m. l. , li h. , li s. , 2010 , , 186 , 308 t. j. , dutta s. n. , di matteo t. , hernquist l. , hopkins p. f. , robertson b. , springel v. , 2006 , , 650 , 791 r. a. , theuns t. , dalla vecchia c. , eke v. r. , frenk c. s. , jenkins a. , kay s. t. , peacock j. a. , pearce f. r. , schaye j. , springel v. , thomas p. a. , white s. d. m. , wiersma r. p. c. , 2009 , , 399 , 1773 d. j. , springel v. , white s. d. m. , de lucia g. , frenk c. s. , gao l. , jenkins a. , kauffmann g. , navarro j. f. , yoshida n. , 2006 , , 365 , 11 l. , dehnen w. , 2010 , , 408 , 669 r. , cen r. , ostriker j. p. , bryan g. l. , hernquist l. , katz n. , weinberg d. h. , norman m. l. , oshea b. , 2001 , , 552 , 473 r. , hernquist l. , katz n. , weinberg d. h. , 1999 , , 511 , 521 t. , colberg j. , springel v. , hernquist l. , sijacki d. , 2008 , , 676 , 33 t. , springel v. , hernquist l. , 2005 , , 433 , 604 j. , kuhlen m. , madau p. , zemp m. , moore b. , potter d. , stadel j. , 2008 , , 454 , 735 k. , borgani s. , murante g. , springel v. , 2009 , , 399 , 497 k. , grasso d. , springel v. , tkachev i. , 2005 , , 1 , 9 k. , jubelgas m. , springel v. , borgani s. , rasia e. , 2004 , , 606 , l97
k. , vazza f. , brunetti g. , tormen g. , 2005 , , 364 , 753 e. , springel v. , hernquist l. , keres d. , 2010 , , 709 , 1138 e. , vogelsberger m. , faucher - giguere c .- a .
, hernquist l. , 2010 , , 725 , 353 e. , vogelsberger m. , hernquist l. , 2011 , in prep a. r. , schaye j. , kay s. t. , dalla vecchia c. , battye r. a. , booth c. m. , 2010 , , 405 , 2161 j. , komatsu e. , nolta m. r. , spergel d. n. , larson d. , hinshaw g. , page l. , bennett c. l. , gold b. , jarosik n. , weiland j. l. , halpern m. , hill r. s. , kogut a. , limon m. , meyer s. s. , tucker g. s. , wollack e. , wright e. l. , 2009 , , 180 , 306 d. j. , hu w. , 1999 , , 511 , 5 c. , lidz a. , zaldarriaga m. , hernquist l. , 2009 , , 703 , 1416 c. s. , white s. d. m. , bode p. , bond j. r. , bryan g. l. , cen r. , couchman h. m. p. , evrard a. e. , gnedin n. , jenkins a. , khokhlov a. m. , klypin a. , navarro j. f. , norman m. l. , et al . , 1999 , , 525 , 554 l. , springel v. , white s. d. m. , 2005 , , 363 , l66 r. a. , monaghan j. j. , 1977 , , 181 , 375 n. y. , 1995 , , 97 , 231 f. , brook c. , mayer l. , brooks a. , rhee g. , wadsley j. , jonsson p. , willman b. , stinson g. , quinn t. , madau p. , 2010
, , 463 , 203 f. , willman b. , mayer l. , brooks a. , stinson g. , valenzuela o. , wadsley j. , quinn t. , 2007 , , 374 , 1479 j. , callegari s. , madau p. , mayer l. , 2011 , arxiv e - prints q. , white s. , boylan - kolchin m. , de lucia g. , kauffmann g. , lemson g. , li c. , springel v. , weinmann s. , 2011 , , p. 164
k. , luki z. , fasel p. , habib s. , warren m. s. , white m. , ahrens j. , ankeny l. , armstrong r. , oshea b. , ricker p. m. , springel v. , stadel j. , trac h. , 2008 , computational science and discovery , 1 , 015003 l. , 1987 , , 64 , 715 l. , katz n. , 1989 , , 70 , 419 l. , katz n. , weinberg d. h. , miralda - escud j. , 1996 , , 457 , l51 s. , springel v. , 2010 , , 406 , 2289 r. w. , eastwood j. w. , 1981 , computer simulation using particles .
mcgraw - hill hopkins p. f. , cox t. j. , dutta s. n. , hernquist l. , kormendy j. , lauer t. r. , 2009 , , 181 , 135 hopkins p. f. , hernquist l. , cox t. j. , dutta s. n. , rothberg b. , 2008 , , 679 , 156 hopkins p. f. , hernquist l. , cox t. j. , keres d. , wuyts s. , 2009 , , 691 , 1424 hopkins p. f. , lauer t. r. , cox t. j. , hernquist l. , kormendy j. , 2009 , , 181 , 486 l. , schmidt w. , niemeyer j. c. , merklein j. , 2011 , , 414 , 2297 s .- i .
, 2002 , journal of computational physics , 179 , 238 a. , frenk c. s. , white s. d. m. , colberg j. m. , cole s. , evrard a. e. , couchman h. m. p. , yoshida n. , 2001 , , 321 , 372 m. , springel v. , dolag k. , 2004 , , 351 , 423 m. , springel v. , enlin t. , pfrommer c. , 2008 , , 481 , 33 n. , hernquist l. , weinberg d. h. , 1992 , , 399 , l109 n. , weinberg d. h. , hernquist l. , 1996 , , 105 , 19 n. , weinberg d. h. , hernquist l. , miralda - escude j. , 1996 , , 457 , l57
g. , white s. d. m. , guiderdoni b. , 1993 , , 264 , 201 jr .
r. c. , 1998 , , 498 , 541 d. , vogelsberger m. , sijacki d. , springel v. , hernquist l. , 2011 , arxiv e - prints a. , trujillo - gomez s. , primack j. , 2010 , arxiv e - prints e. , et al . , 2011 , , 192 , 18 l. b. , 1977 , , 82 , 1013 j. c. , hernquist l. , 1994 , , 437 , 611 j. c. , hernquist l. , 1996 , , 464 , 641 n. l. , mccarthy i. g. , bower r. g. , theuns t. , crain r. a. , 2009a , , 395 , 180 n. l. , mccarthy i. g. , bower r. g. , theuns t. , crain r. a. , 2009b , , 395 , 180 j. j. , 1992 , , 30 , 543 j. j. , 2005 , reports on progress in physics , 68 , 1703 b. , ghigna s. , governato f. , lake g. , quinn t. , stadel j. , tozzi p. , 1999
, , 524 , l19 b. , governato f. , quinn t. , stadel j. , lake g. , 1998 , , 499 , l5 j. , 1997 , journal of computational physics , 136 , 41 d. , springel v. , marcus r. , vogelsberger m. , hernquist l. , 2012 , arxiv e - prints c. , katz n. , hernquist l. , weinberg d. h. , dav r. , 2002 , , 571 , 1 g. , borgani s. , brunino r. , cha s .- h .
, 2011 , , 417 , 136 t. , burkert a. , 2003 , , 597 , 893 k. , cen r. , hernquist l. , ostriker j. p. , springel v. , 2005 , , 627 , 608 j. f. , frenk c. s. , white s. d. m. , 1996 , , 462 , 563 j. f. , frenk c. s. , white s. d. m. , 1997 , , 490 , 493 j. f. , ludlow a. , springel v. , wang j. , vogelsberger m. , white s. d. m. , jenkins a. , frenk c. s. , helmi a. , 2010 , , 402 , 21 b. w. , bryan g. , bordner j. , norman m. l. , abel t. , harkness r. , kritsuk a. , 2004 , arxiv astrophysics
e - prints b. w. , nagamine k. , springel v. , hernquist l. , norman m. l. , 2005 , , 160 , 1 f. r. , jenkins a. , frenk c. s. , colberg j. m. , white s. d. m. , thomas p. a. , couchman h. m. p. , peacock j. a. , efstathiou g. , the virgo consortium 1999 , , 521 , l99
u. , 1998 , , 115 , 19 w. j. , et al .
, 2010 , , 401 , 2148 s. , et al . , 1999 , , 517 , 565 m. , springel v. , 2010 , , p. 1851 c. , springel v. , enlin t. a. , jubelgas m. , 2006 , , 367 , 113 w. h. , schechter p. , 1974
, , 187 , 425 d. j. , 2008 , journal of computational physics , 227 , 10040 d. j. , 2012 , journal of computational physics , 231 , 759 e. , sijacki d. , springel v. , 2008 , , 687 , l53 f. a. , 2000 , progress of theoretical physics supplement , 138 , 609 j. i. , hayfield t. , 2011 , arxiv e - prints j. i. , hayfield t. , agertz o. , 2010 , , 405 , 1513 j. i. , lake g. , agertz o. , debattista v. p. , 2008
, , 389 , 1041 m. j. , ostriker j. p. , 1977
, , 179 , 541 j. a. , haehnelt m. g. , viel m. , 2007 , , 374 , 196 a. g. , filippenko a. v. , li w. , treffers r. r. , schmidt b. p. , qiu y. , hu j. , armstrong m. , faranda c. , thouvenot e. , buil c. , 1999 , , 118 , 2675 b. , yoshida n. , springel v. , hernquist l. , 2004 , , 606 , 32 m. , monaghan j. j. , 2011 , arxiv e - prints c. , et al . ,
2011 , arxiv e - prints c. , tissera p. b. , white s. d. m. , springel v. , 2008 , , 389 , 1137 j. , dalla vecchia c. , booth c. m. , wiersma r. p. c. , theuns t. , haas m. r. , bertone s. , duffy a. r. , mccarthy i. g. , van de voort f. , 2010 , , 402 , 1536 p. , finoguenov a. , miniati f. , bhringer h. , briel u. g. , 2004 , , 426 , 387 i. , schmitt d. , 1981 , , 97 , 373 d. , springel v. , di matteo t. , hernquist l. , 2007 , , 380 , 877 d. , vogelsberger m. , keres d. , springel v. , hernquist l. , 2011 , arxiv e - prints j. , 1977 , , 211 , 638 b. , sigurdsson s. , abel t. , 2008 , , 385 , 1443 v. , 2005 , , 364 , 1105 v. , 2010a , , 401 , 791 v. , 2010b , , 48 , 391 v. , di matteo t. , hernquist l. , 2005 , , 361 , 776 v. , hernquist l. , 2002 , , 333 , 649 v. , hernquist l. , 2003a , , 339 , 289 v. , hernquist l. , 2003b , , 339 , 312 v. , hernquist l. , 2005 , , 622 , l9 v. , wang j. , vogelsberger m. , ludlow a. , jenkins a. , helmi a. , navarro j. f. , frenk c. s. , white s. d. m. , 2008 , , 391 , 1685 v. , white s. d. m. , jenkins a. , frenk c. s. , yoshida n. , gao l. , navarro j. , thacker r. , croton d. , helly j. , peacock j. a. , cole s. , thomas p. , couchman h. , evrard a. , colberg j. , pearce f. , 2005 , , 435 , 629 j. m. , norman m. l. , 1992 , , 80 , 753 e. j. , brunino r. , mitchell n. l. , michielsen d. , hopton s. , pearce f. r. , bryan g. l. , theuns t. , 2008 , , 390 , 1267 r. , 2002 , , 385 , 337 a. , toomre j. , 1972 , , 178 , 623 p. , vogelsberger m. , sijacki d. , springel v. , hernquist l. , 2011 , arxiv e - prints h. , sills a. , pen u. , 2007 , , 377 , 997 r. , 2011 , , 526 , a158 f. , brunetti g. , kritsuk a. , wagner r. , gheller c. , norman m. , 2009 , , 504 , 33 m. , bolton j. s. , haehnelt m. g. , 2009 , , 399 , l39 m. , lesgourgues j. , haehnelt m. g. , matarrese s. , riotto a. , 2005 , , 71 , 063534 m. , helmi a. , springel v. , white s. d. m. , wang j. , frenk c. s. , jenkins a. , ludlow a. , navarro j. f. , 2009 , , 395 , 797 m. , white s. d. m. , 2011a , , p. 168
m. , white s. d. m. , 2011b , , 413 , 1419 m. , white s. d. m. , helmi a. , springel v. , 2008 , , 385 , 236 j. w. , veeravalli g. , couchman h. m. p. , 2008
, , 387 , 427 d. h. , dav r. , katz n. , hernquist l. , 2004 , , 601 , 1 s. d. m. , frenk c. s. , 1991 , , 379 , 52 s. d. m. , rees m. j. , 1978 , , 183 , 341 r. p. c. , schaye j. , theuns t. , dalla vecchia c. , tornatore l. , 2009 , , 399 , 574 y. b. , 1970 , , 5 , 84
for the different arepo runs . the black line shows the characteristics for a mesh where @xmath95 vertex points are randomly distributed .
our implemented regularisation scheme tries to keep the maximum face angle small to avoid very distorted voronoi cells .
this can clearly be seen in the upper left panel , where the random distribution deviates significantly from the regularised simulation distributions .
the regularisation also affects the offset between geometrical centre of the cell and the vertex location shown in the bottom left panel .
it is important to avoid too large offsets here since they introduce inaccuracies in the linear reconstruction step of the muscl - hancock scheme .
the `` roundness '' of the cells is measured by @xmath152 , which is closer to @xmath153 ( the value of a sphere ) for the regularised simulations .
again , this is a measure of the regularity of the voronoi mesh . for each face of a voronoi cell
, arepo needs to solve a riemann problem .
since each face is only solved once , the average number of riemann problems per cell is given by half the average number of its faces .
the distribution of voronoi faces per cell is shown in the top right panel.,scaledwidth=48.0% ] .
our re-/derefinement scheme guarantees that the mass of a voronoi cell never deviates by more than approximately a factor of 2 from the target gas mass , which is in our case chosen to be the mean baryon mass of a cell for a uniform distribution assuming all cells have equal volume .
bottom panel : distribution of masses of the stellar particles divided by the imposed target gas mass for the voronoi cells .
our constraints for the mass of cells and our scheme for creating star particles result in a quite narrow mass distribution of stellar particles , where each stellar particle deviates at most a factor of @xmath154 from the desired target gas mass , except for a tiny number of lighter star particles , whose mass is however always larger than a quarter of the target gas mass . ,
title="fig:",scaledwidth=45.0% ] .
our re-/derefinement scheme guarantees that the mass of a voronoi cell never deviates by more than approximately a factor of 2 from the target gas mass , which is in our case chosen to be the mean baryon mass of a cell for a uniform distribution assuming all cells have equal volume .
bottom panel : distribution of masses of the stellar particles divided by the imposed target gas mass for the voronoi cells .
our constraints for the mass of cells and our scheme for creating star particles result in a quite narrow mass distribution of stellar particles , where each stellar particle deviates at most a factor of @xmath154 from the desired target gas mass , except for a tiny number of lighter star particles , whose mass is however always larger than a quarter of the target gas mass .
, title="fig:",scaledwidth=45.0% ] here we present some basic geometric properties of the voronoi mesh in our cosmological simulations . as discussed in section 2 , it is beneficial for the accuracy of arepo for the voronoi mesh to remain as regular as possible throughout the simulation . to this end ,
the motion of the mesh vertices is slightly modified for highly distorted cells in order to obtain a mesh that has more `` roundish '' cells and is closer to a centroidal voronoi tessellation where geometrical the centres of the cells coincide with the vertex points of the individual voronoi cells .
we can quantify the quality of a mesh in a similar way as was done in s10 .
there , mesh quality indicators were calculated for a non - radiative simulation of a massive cluster ( the santa barbara cluster , * ? ? ?
* ) , and it was demonstrated that the mesh indeed preserved its desired properties of reasonably round cells during the simulation .
however , in this paper we have used a modified regularisation scheme based on the maximum face - angle , and we include cooling , star formation , and feedback , which drastically increases the dynamic range the simulations need to address .
it is therefore important to verify whether our regularisation scheme performs sufficiently well in these more demanding simulations . in fig .
[ fig : cell_shape_statistics ] , we first show in the top left panel the maximum face angle distribution at @xmath5 for all arepo simulations .
our new regularisation scheme should guarantee that these angles do not get too large .
we have set up the regularisation such that the mesh is steered towards face angles smaller than @xmath10 . as the upper left panel of fig .
[ fig : cell_shape_statistics ] demonstrates , the mesh indeed avoids very large face angles with a maximum in the distribution around the desired value , i.e. the regularisation works as expected .
the remaining three panels of fig .
[ fig : cell_shape_statistics ] show distribution functions for the number of faces of the voronoi cells , for the distance @xmath155 of mesh - generating points to the geometric centres of their cell in units of the fiducial cell radius @xmath156 of each cell , and finally , for the @xmath157 parameter which measures how `` roundish '' each cell is ( here @xmath158 is the total surface area of a cell ) .
the corresponding distributions for a random poisson sample of @xmath95 vertex points are shown as black lines in all panels , for comparison .
note that we do not explicitly control the offset @xmath155 in our regularisation scheme , but the face angle criterion is of course correlated with this quantity .
this allows us to obtain also quite small values for the distance of the geometric centre of the cell from the location of the mesh generating points .
this is important for the numerical accuracy of the linear reconstruction .
overall , our voronoi mesh is significantly `` rounder '' and much closer to a centroidal voronoi tessellation than the mesh of a random point distribution .
the goal of our re- and derefinement strategy is to keep all cell masses close to a predefined target cell mass , which we have chosen to be the total baryonic mass in the box divided by the total number of cells at the initial time . in this way , a narrow mass spectrum for the formed stellar particles is obtained , and a straightforward and even - handed comparison to the sph calculations done with gadget becomes possible . as we have described in section 2 , we guarantee the approximately constant mass resolution by splitting cells that reach a mass larger than @xmath14 into two cells , while cells dropping in mass below @xmath159 are dissolved .
we note that thanks to the lagrangian mesh motion in arepo , these re-/derefinement operations are invoked only rarely . in the top panel of fig .
[ fig : cell_mass_statistics ] we demonstrate that the re- and derefinement scheme successfully keeps the cell masses in the desired mass range , yielding approximately a @xmath160-normal distribution around the target mass within the desired bounds .
the mass distribution of stellar particles at @xmath5 is shown in the bottom panel of fig .
[ fig : cell_mass_statistics ] .
since our star formation implementation does not allow star particles to be formed with a mass larger than @xmath161 , the stellar particle mass distribution is cut off at that value . on the other hand ,
the derefinement operations are constrained by the requirement that neighbouring cells can not both be derefined in the same time step . as a result , it is possible that some cells can temporarily have masses below @xmath19 for a few time steps , and hence also some star particles with masses smaller than this value may form in principle . to protect against the formation of unreasonably low mass star particles ,
which may be prone to two - body effects , we however do not allow cells with mass less than @xmath17 to form any star particles , imposing a lower limit in the star particle mass distribution . as the bottom panel of fig .
[ fig : cell_mass_statistics ] shows , star particles with masses between @xmath17 and @xmath19 make up only a tiny fraction , as desired , such that the suppression of the formation of extremely small star particles does not lead to any appreciable error . note that star - forming cells with a mass above @xmath162 will form stars with exactly @xmath12 , keeping the rest of the gas mass in the cell
this explains the small spike in the distribution at that mass value .
our analysis thus confirms that star particles with a narrow range of masses are formed , which helps to limit two - body relaxation effects and is well matched to the fixed gravitational softening we use for collisionless particles .
the arepo simulations presented in this paper use an adaptive gravitational softening length with a lower floor for the gas , whereas the sph simulations done with gadget employ a fixed comoving softening length equal to the lower floor .
we checked that this does not bias our results in any way . to demonstrate this point we show in fig .
[ fig : sfr_hist_appendix ] the star formation history of the standard a_l20n128 simulation together with a run where we held the gravitational softening of the cells fixed at the same value as in the sph calculations .
clearly this leads only to very minor differences in the star formation history .
we also checked that disk half mass radii are not affected by this .
as stated above this is due to the fact that star forming gas is pressurised by an effective eos of the ism and therefore gravitational softening effects do not play an important role in that regime .
the mesh regularisation in arepo can be done in different ways and we have used a scheme that is based on the maximum face angle as described above .
we have also done simulations with the original regularisation scheme presented in @xcite , which is based on the displacement of the geometric centre of the cell .
the resulting star formation history for the a_l20n128 is also shown in fig . [
fig : sfr_hist_appendix ] .
again this leads only to minor changes .
we note that this difference becomes smaller with resolution and essentially vanishes .
regularisation scheme differences are only relevant for very low resolution simulations .
disk sizes are also not significantly affected by the details of the regularisation scheme . in fig .
[ fig : gas_map_regul ] , we show gas density maps of a matched object in a_l20n128 and the a_l20n128 run , which features an alternative regularisation scheme .
we stress again that l20n128 is our lowest resolution setup , and therefore most sensitive to details of how exactly the mesh is treated .
but even in this regime the differences in the gas density maps are very small , and the size and overall extent of the gas disk do not change . note that due to the low resolution of l20n128 , individual voronoi cells are clearly visible in the map , highlighting the changes in mesh geometry and cell shapes induced by different regularisation schemes .
the total runtime of our arepo simulations is in the worst case only about @xmath143 longer than the corresponding gadget run at the same nominal resolution .
[ fig : runtime_stats ] shows in which code parts most of this cpu time is spent .
obviously , the rather complex mesh construction and mesh update in arepo takes up a significant amount of time .
however , in both codes the tree - pm based gravity calculation consumes a large amount of time , too , alleviating speed differences in the hydrodynamical solvers . in any case , we argue that arepo is actually surprisingly fast given the extremely complex mesh management operations that are required . given the small difference in raw speed , one can rightly describe the overall performance of both codes as similar .
however , accounting for the fact that we here find that sph leads to a systematic and significant offset in the cooling rates of haloes at late times , as well as in galaxy sizes , the discussion of cpu time requirements at the same nominal resolution is a bit moot .
ideally one would like to select the fastest method for a given desired accuracy , and as it appears , this can be reached with arepo more efficiently , whereas standard sph s systematic bias may be difficult or even impossible to overcome .
the white parts in fig . [ fig : runtime_stats ] account for the time spent in the time integration , star formation , cooling , and other minor contributions .
we note that the i / o contribution is a bit larger than typical , because we wrote out snapshots at a very high frequency .
and @xmath42 cores . in all runs an identical simulation size of @xmath84 in a @xmath163 box was used , with a @xmath164 fft for the long - range gravity calculation .
the black solid line shows the total wall - clock time for a full step as a function of the number of cores used , averaged over three steps at high redshift .
the code was run with all physics enabled , and all code overhead was included in these averages .
the other measurements included in the figure give the times for the most important individual parts of the code , which are the voronoi mesh construction , the short - range and long - range gravity calculations , the hydrodynamical flux calculation , the domain decomposition and the tree construction .
the black dashed line indicates the ideal strong scaling .
, scaledwidth=48.0% ] in these tests .
we show in black the total times averaged for three full steps , and with different colours the most important code parts , as labelled .
the top panel refers to a series with `` fully loaded '' nodes , where we use a load that is close to the maximum we can fit into the memory available per ranger core . in the bottom panel , we have reduced the load by a factor of 2 , allowing us to extend the tests to a core count of 6912 ( for such large mpi partition sizes , less application memory remains available on the specific machine ) . here
, however , the communication costs in three parts of our code become quite substantial and lead to a noticeable negative impact of the scalability .
the particle load of the top panel goes from @xmath165 to @xmath166 , whereas for the bottom panel it goes from @xmath167 to @xmath166 .
the black dashed line indicates the ideal weak scaling .
, scaledwidth=48.0% ] in fig .
[ fig : strongscaling ] , we show a strong scaling test of arepo , from @xmath168 to @xmath42 cores , carried out on ranger at the texas advanced computing center ( tacc ) . here , the simulation size has been kept constant at @xmath84 , and only the number of compute cores has been increased .
the reported times have been averaged for three full time - steps at high redshift . in this regime ,
gravity accounts for @xmath169 of the computational time , mesh - construction and mesh bookkeeping consumes @xmath170 of the time , while the calculation of the hydrodynamical fluxes amounts to @xmath171 of the time .
the remainder is needed for miscellaneous items such as domain decomposition , radiative cooling and star formation . in order to clearly show the scalability of the most important different parts of our code
, we have included separate measurements in fig .
[ fig : strongscaling ] for long - range gravity ( by means of ffts ) , short - range gravity ( done through a tree - walk ) , the necessary tree construction , the mesh construction and management , the hydrodynamic flux calculations , and the domain decomposition .
we see that the code shows quite good strong scaling , despite the tightly coupled nature of the system .
some losses are in particular apparent for the mesh construction .
these arise because the more cores are used , the larger the number of spatial domains in which the simulation volume is cut .
this enlarges the surface area of domain boundaries , and as a result the cumulative `` ghost '' region volume in which the mesh has to be constructed twice on two neighbouring processes to ensure seamless consistency across the domain boundary . for large - scale cosmological applications it is more important that the code parallelisation shows good weak scaling behaviour .
gadget has been used for many large - scale simulations and we will demonstrate here that arepo performs also well in weak scaling tests . for these tests the computational load per core is kept constant , and the simulation volume and particle / cell number is scaled accordingly . in fig . [
fig : weakscaling ] , we show two different weak scaling series carried out on ranger at the tacc . they differ in their particle and cell load per core . in the results shown in the top panel , this load was `
maximal ' , corresponding to @xmath172 million cells and particles per core , where the current version of arepo requires up to 1600 mb memory consumption per core in the peak .
because some memory is needed for the operating system and the mpi communication library , not all of the physical memory is available for the application code on the ranger platform .
unfortunately , the amount of memory consumed by the mpi subsystem increases with increasing size of the mpi partition .
this in fact prevents us from running the `` full load '' configuration for partitions larger than 4096 cores .
this also prompted us to create a second weak scaling series shown in the bottom panel of fig .
[ fig : weakscaling ] , where we reduced the load by almost a factor of 2 .
this allows us to scale arepo up to @xmath173 cores with @xmath173 mpi tasks .
it is apparent that the short - range gravitational tree calculation , the voronoi mesh construction , and the hydrodynamical flux calculation show excellent weak scaling ( which runs horizontally in the plots of fig .
[ fig : weakscaling ] ) . however , there are deviations from perfect scalability for the fft - based long - range gravitational force calculation , the domain decomposition , and the tree construction .
this is primarily because all three of these parts involve substantial all - to - all communication . for larger processor counts , especially for @xmath173 cores ,
these communication costs start to affect the weak scalability of our code and lead to losses in efficiency .
one reason for the sub - optimal scaling in this regime lies in the fft part of the code .
the slab - based parallelisation of the ffts we use for the long - range gravity calculation does not scale to core numbers larger than the size of the fft , a regime we enter once more than 4096 cores are used ( because the fft - size reaches @xmath174 at this point and can not be grown further due to memory constraints ) .
we note that this can easily be optimised by either using a block - structured fft decomposition or by using a mixed approach for the parallelisation with either threads or openmp .
overall we find that arepo shows good weak scaling behaviour and can readily be applied to large - scale cosmological hydrodynamics simulations . | we present the first hydrodynamical simulations of structure formation using the new moving mesh code arepo and compare the results with gadget simulations based on a traditional smoothed particle hydrodynamics ( sph ) technique .
the two codes share the same tree - pm gravity solver and include identical sub - resolution physics for the treatment of star formation , but employ different methods to solve the equations of hydrodynamics .
this allows us to assess the impact of hydro - solver uncertainties on the results of cosmological studies of galaxy formation . in this paper
, we focus on predictions for global baryon statistics , such as the cosmic star formation rate density , after we introduce our simulation suite and numerical methods
. properties of individual galaxies and haloes are examined by @xcite , while a third paper by @xcite uses idealised simulations to analyse in more detail the differences between the hydrodynamical schemes .
we find that the global baryon statistics differ significantly between the two simulation approaches .
arepo shows systematically higher star formation rates at late times , lower mean temperatures averaged over the simulation volume , and different gas mass fractions in characteristic phases of the intergalactic medium , in particular a reduced amount of hot gas .
although both hydrodynamical codes use the same implementation of cooling and yield an identical dark matter halo mass function , more gas cools out of haloes in arepo compared with gadget towards low redshifts , which results in corresponding differences in the late - time star formation rates of galaxies .
we show that this difference is caused by a higher heating rate with sph in the outer parts of haloes , owing to viscous dissipation of sph s inherent sonic velocity noise and sph s efficient damping of subsonic turbulence injected in the halo infall region , and because of a higher efficiency of gas stripping in arepo . as a result of such differences
, arepo leads also to more disk - like morphologies in the moving mesh calculation compared to gadget .
our results hence indicate that inaccuracies in hydrodynamic solvers can lead to comparatively large systematic differences even at the level of predictions for the global state of baryons in the universe .
[ firstpage ] cosmology : galaxy formation cosmology : dark matter methods : numerical |
50% of the traffic in cellular networks today is video @xcite .
mounting evidence suggests that video will keep increasing its share of the cellular traffic at an even faster pace @xcite .
the reason behind this phenomenon is the explosive demand for high quality video streaming from mobile devices ( e.g. , tablets , smart - phones ) .
the challenge for mobile network operators ( mnos ) is to offer higher data rates that can keep up with this demand for high quality video .
heterogenous cellular networks ( hcns ) , illustrated in fig . [ fig : system - model ] , are envisioned to be one of the solutions to this problem .
hcns introduce low power base stations ( bs ) like pico bs ( pbs ) and femto bs ( fbs ) , that form around them picocells and femtocells respectively . lower transmission power from these small cells
reduces the transmission range and allows improved spatial reuse .
hence , the first novel feature of hcns is the _ higher wireless capacity _ they offer to the complete macrocell .
the second novel feature of hcns is that they can _ lower the use of the backhaul capacity _ by employing caching at the small cell bss @xcite .
caching enables local access of frequently requested videos and this means lower utilization of the backhaul links between a bs and the video server .
these two features of hcns constitute them a central component of the envisioned 5 g cellular network architecture .
research for video distribution in hcns has focused primarily on caching , with the objective to reduce the startup playback delay of the video for each user @xcite , or lower the costs for the operator @xcite . in this paper
we are concerned with the first novel feature of hcns which is the higher wireless capacity .
we focus on this topic since hcns introduce a new way for sharing the wireless resources . with the time domain resource partitioning ( tdrp ) mechanism @xcite
, the mbs shuts off its transmissions for a fraction @xmath0 of the available resources during which the small cells can achieve a higher data rate ( fig .
[ fig : problem - model ] ) . during the fraction @xmath1
, there is _ intra - cell _ interference since the mbs transmits simultaneously with the small cells .
this technique was recently standardized through the concept of almost blank subframes ( abs ) and regular subframes ( rs ) in 3gpp lte - a under the more general name of enhanced inter - cell interference coordination ( eicic ) @xcite .
one important detail is that the lte - a standard currently allows the dynamic adaptation of @xmath0 but it does not specify how it should be configured . given the increasing number of video streaming users in cellular networks , and the necessity of tdrp , it is of outmost importance to perform optimally both the configuration of @xmath0 and the allocation of the wireless resources in an hcn .
hence , the specific questions that should be answered in this case are : what is the optimal @xmath0 when we have video traffic ? what is the best video quality that each user should receive ? what happens when a subset of the users receive video
? currently there are no definite answers to these pressing questions .
* related work . *
tdrp for hcns is a topic investigated only recently because abs / rs were also very recently standardized in lte - a .
the authors in @xcite derived the optimal fraction from the available abs and rs resources that each user should be allocated ( a representative rate allocation is illustrated in fig . [ fig : problem - model ] ) under a proportionally fair rate allocation ( pfra ) metric .
the authors of that work assumed a constant fraction of abs @xmath0 that is configured by the hcn operator .
in another recent work reported in @xcite , the authors investigated the joint optimization of tdrp and user association ( for traffic offloading ) but with an assumption for equal rate allocation to the associated users . to the best of our knowledge
there is no work that addresses tdrp in hcns for video distribution and streaming .
as we already discussed , much of the early research work for video streaming hcns has focused on caching @xcite , or exploiting particular features like the density of the small cells @xcite . however , these works assumed the availability of a constant fraction of the resources for the mbs and the small cells that is effectively translated to a constant @xmath0 . on the other hand ,
multi - user rate allocation for video streaming has been a topic thoroughly investigated the last few years for specific types of wireless networks .
the works were primarily motivated from the network utility maximization ( num ) framework @xcite . from the category of works that were based on num , the ones that are more related to this paper focused on cellular networks and considered more details of the physical layer ( phy ) .
for example the authors in @xcite investigated scheduling and resource allocation for a downlink lte system that employs discrete decisions for optimizing the selected video streaming quality . the same problem , but for scalable encoded video
, was considered in @xcite .
another class of works focused on optimizing rate / resource allocation with the objective to improve the playback performance of video clients @xcite . in the last works the authors take into consideration recent standardization developments in streaming and in particular dynamic adaptive streaming over http ( dash ) .
however , these works do not target hcns and assume access to fixed capacity resources .
* contributions .
* in this paper , we present contributions on three fronts .
first , we present a comprehensive _ joint tdrp , rate allocation , and video quality selection ( jtravqs ) _ optimization framework for video streaming in hcns .
the framework identifies the optimal tdrp @xmath0 , the rate allocated to each user , and the video quality description for each user , so as to maximize the aggregate video quality in the hcn .
our framework includes additional system - level parameters like the fraction of users that receive video .
the np - hard problem is solved with a primal - dual approximation algorithm that provides an asymptotically optimal solution .
our solution approach decomposes the problem into simpler subproblems , making them amenable to fast well - known solution algorithms .
second , we propose enhancements to the basic optimization framework that allow it to support dash - based video streaming . additional system parameters like the buffer contents of individual users , time - dependent user population , and channel capacity with tcp , are taken into account to optimize the the playback performance @xcite .
third , we present a thorough performance evaluation our main framework against : a ) a video - unaware system that jointly optimizes the tdrp and rate allocation under a pfra metric @xcite .
b ) from the results of our scheme and that of pfra , we can infer the performance of a system that applies optimized rate allocation and video quality selection ( ravqs ) but with a fixed tdrp @xcite .
finally , our enhanced system for dash , that considers the content of the playback buffer , is compared against a buffer - aware system that again uses fixed resources .
* main results . *
our results reveal that : i ) for video streaming tdrp should be more aggressive in favor of the small cells when compared to tdrp optimization under a pfra metric . in particular even for 4 small cells and 100 users , the optimal @xmath0 should be nearly 22.2% higher than the optimal @xmath0 under pfra .
video quality is improved by a factor of 50%-70% for this scenario .
ii ) using a fixed _ but still optimal tdrp under a pfra metric _ , and then performing a ravqs optimization as an afterthought , is still suboptimal .
in particular for a population of 100 users and 4 small cells , the previous approach leads to an average video quality loss of 18.6% when compared to our approach .
iii ) for a dash - based system with a fixed , _ but again optimal tdrp under pfra _ , our optimization has more significant impact .
in particular the rebuffering time / events of the clients can be reduced by more than 50% for a static network and 60% for a network with user churn . * paper organization . * the rest of the paper is organized in the following sections .
section [ section : system - model ] describes in detail the system model . in section [ section : optimization ]
we present the formulation and the solution of the optimization problem we introduce in this paper , while its extension for dash is presented in section [ section : optimization - dash ] .
performance evaluation results are presented in section [ section : performance - evaluation ] , and finally we conclude in section [ section : conclusions ] .
* network model . * in fig . [ fig : system - model ] we present the network that we study in this paper and it includes a single macrocell with a mbs , the pbss , and the users . each bs @xmath2 in the set @xmath3 communicates with the set of associated users @xmath4 .
we also denote with @xmath5 a subset of the users associated to bs @xmath2 that are not optimized in a video - aware fashion .
this parameter allows us to investigate the possibility that a fraction @xmath6 of the users are optimized . during the fraction @xmath0 of the abs resources
all the small cells transmit and interfere with every active user in the network .
thus , we consider _ resource reuse _ across bss of the same tier ( pbss in our case ) which is one of the main benefits of small cells since it allows spatial reuse .
the aggregate average interference power that user @xmath7 receives is denoted as @xmath8 . during the fraction of the non - blanked resources , or regular subframes , @xmath1 both the mbs and pbss transmit and the aggregate interference power that a user receives is denoted as @xmath9 .
* user model .
* each bs @xmath2 transmits with unicast streaming video @xmath7 to the similarly denoted user .
the users associate to a bs by using an signal - to - interference plus noise ratio ( sinr ) biasing rule @xcite , i.e. , a user is associated to the small cell @xmath2 , and not the mbs , if the following is true : @xmath10 .
this ensures that users are offloaded to the small cells @xcite .
our primary objective is a static user population similarly to the literature @xcite , since we are interested to optimize the system operation within the complete playback duration of the video .
however , motivated by recent experimental results that identify slow user variations in the cell throughout the day @xcite , we also evaluate our system for this more dynamic scenario . * video streaming and playback model .
* without loosing generality we assume that all the bss are assumed to have cached the videos for all the users @xcite .
now if the video representation that is transmitted to user @xmath7 is indexed by @xmath11 , the average bitrate that must be sustained is @xmath12 where @xmath13 is the size of the @xmath11-th video representation , @xmath14 is the total playback time of the video and @xmath15 is the duration of playable content received during startup buffering ( time 0 ) .
this formulation ensures that the average probability of rebuffering events is zero @xcite . in the first part of our optimization in section [ section : optimization ] , this is the condition we adopt since we are interested to reach a decision once for the duration of the video streaming session .
however , with dash this formula is revised in section [ section : optimization - dash ] .
* channel & phy model .
* nodes use a single omni - directional half - duplex antenna .
the channel from the @xmath2-th bs to the @xmath7-th user is denoted as @xmath16 .
the fading coefficients are independent and @xmath16@xmath17@xmath18 , i.e. , they are complex gaussian random variables with unit variance and mean equal to @xmath19 that depends on path loss and shadowing effects according to the lte channel model @xcite .
all the channels are considered to be block - fading rayleigh and quasi - stationary , that is they remain constant for the coherence period of the channel that is equal to the transmission length of the complete phy block .
additive white gaussian noise ( awgn ) is assumed at every receiver with variance @xmath20 .
the transmission power that the pbs and mbs use is @xmath21 , and @xmath22 respectively .
* mcs & cqi . * a modulation and coding scheme ( mcs ) with @xmath23 bits / symbol is used by each bs while its value is determined by each pbs independently and optimally as we will later explain .
the set of available mcss is @xmath24 , i.e. , we assume that the most spectral efficient quadrature amplitude modulation ( qam ) mcs is 128-qam .
we also assume that users provide only _ average channel quality indicator _ ( cqi )
feedback to the bss .
modeling the quality - of - experience ( qoe ) of users in video streaming applications is not easy .
qoe is affected both by the video signal quality and delay @xcite . in this subsection
, we define a utility model only for the quality of the video signal while during the analysis of our optimization framework we discuss our approach for minimizing the effects of delay .
the main objective of our video quality model is to capture the rate - distortion ( rd ) relationship of different representations of each video stream .
this will allow our optimization framework to allocate resources to videos depending on their quality . in this paper
we assume we have the rd information information for each frame @xmath25 that belongs to representation @xmath11 of video @xmath7 and consists of its size @xmath26 in bits and the importance of the frame for the overall reconstruction quality of the video denoted as @xmath27 @xcite . in practice
, @xmath27 is the total decrease in the mean square error ( mse ) distortion that will affect the video if the frame is decoded by the video player @xcite .
the value of the mse in @xmath27 includes both the distortion that is added when frame @xmath25 is not decoded , and also the frames that have a decoding dependency with @xmath25 .
for an i frame includes the @xmath28 of the p and b frames that depend on it . ] hence , the video quality model considers also the possible drift that might occur due to the inability to decode a particular video frame .
these values can be obtained easily but only during the offline encoding of the video as discussed in @xcite . consequently , the aggregate video quality of a group of video frames indexed by @xmath29 ( also referred to as segment to ensure consistency with dash terminology ) , that belong to representation @xmath11 of video @xmath7 , is the average : @xmath30 this fraction is the average mse reduction of the frames contained in a dash segment or packet , versus their total number . _
this formulation is in line with our initial objective since it expresses the `` value '' for a group of frames .
_ for a group of segments starting from @xmath31-th segment until the end of the video , we can similarly characterize the video quality as : @xmath32 for packet - based video , this rd information associated with a packet can be contained in each packet header . in the case of scalable video the information about the importance of a packet is already embedded in the header since it indicates the video layer that the packet belongs . for segment - based dash streaming a media presentation description ( mpd ) file is already used for conveying a subset of this information @xcite .
hence , the model can support packetized non - scalable , scalable , and segment - based video .
the final result of the previous discussion is that a single video for user @xmath7 will be available at the following discrete set of qualities @xmath33 indicating the set of available representations for each user / video @xmath7 .
it is important to understand the use of the previous model in our optimization . in our initial framework , where the problem is solved for the complete playback duration of the video , the formulation in is used by setting @xmath31=0 , i.e. , we use the average quality of the complete video .
however , for dash the optimization is solved during a specific time period @xmath31 and captures the video quality of the remaining segments that still need to be communicated . to complete our discussion
, we have to recall that our optimization targets a heterogeneous user population where a subset of them do not receive video .
when we have elastic flows , or when the users do not participate in the video - aware optimization , then rate allocation is exercised with a pfra metric @xcite , i.e. , @xmath34 is generated by taking the logarithm of the communication rate achievable by user @xmath7 .
we consider that the bss optimize independently the phy parameters of the point - to - point links , as it is typically done in wireless communication systems @xcite . to estimate the average communication rate that each user @xmath7 achieves when it is associated to bs @xmath2 we proceed as follows .
the bs receives from each user @xmath7 the average channel gain @xmath35 $ ] , and also the local estimate of the interference power @xmath8 , @xmath9 .
note that the average channel gains , or the cqi in the lte terminology , mentioned above can be transmitted from each user to the bs with low network overhead since they only correspond to path loss and shadowing .
hence , during an abs , since a user receives the aggregate interference from all the simultaneously transmitting pbss , the average sinr between the pbs and user @xmath7 is : @xmath36=\frac{p_\text{pbs}\operatorname{\mathbb{e}}[|h_{\text{pbs},i}|^2]}{i_{\text{abs},i}+\sigma^2}\ ] ] during the rs the mbs is also active and so the sinr of users associated to a pbs and the mbs are @xmath37=\frac{p_\text{pbs}\operatorname{\mathbb{e } } [ |h_{\text{pbs},i}|^2 ] } { i_{\text{rs},i}+\sigma^2},~\text{and}~\operatorname{\mathbb{e } } [ \gamma^{\text{rs}}_{i } ] = \frac{p_\text{mbs } \operatorname{\mathbb{e } } [ |h_{\text{mbs},i}|^2 ] } { i_{\text{rs},i}+\sigma^2},\ ] ] respectively .
the average sinr expressions allow each bs @xmath2 to estimate the resulting average data rate for each associated user @xmath7 under mcs @xmath23 as : @xmath38 where @xmath39 is the symbol rate , @xmath40 is the efficiency of the mcs , and the probability of symbol error @xmath41 under @xmath42-qam is @xcite : @xmath43 } \big ) \end{aligned}\ ] ] in our system the phy and link - layer system at each bs selects optimally for the average sinr the mcs that ensures the highest point to point communication rate @xcite .
this is formally written as : @xmath44
now we are ready to define formally the problem we address in this paper . for
each user @xmath7 associated to bs @xmath2 the hcn must select the video representation with the highest quality , and the rate allocated to it . for the complete hcn
the globally optimal tdrp must be calculated .
first we define the optimization variables .
let @xmath45 indicate whether user @xmath46 is served with video representation @xmath11 in an abs and rs respectively .
let also @xmath47 $ ] denote the fraction of the abs resources that the pbs allocates to @xmath48 for streaming the video representation @xmath11 .
similarly for the rs , we define @xmath49 $ ] .
hence , the decisions of each bs @xmath2 are : ( a ) the _ video quality selection ( vqs ) _ vector for all the associated users , i.e. , @xmath50 , and ( b ) the _ rate allocation ( ra ) _ vector for all users , i.e. , @xmath51 .
similarly the vqs and ra vectors for the regular slots . also the global resource partitioning decision @xmath0 . to minimize the notation later in our solution
, we also define different concatenations of the variable vectors as follows : @xmath52 , @xmath53 , and similarly for @xmath54 .
the objective for the hcn operator is to maximize the average aggregate delivered video quality captured by : @xmath55 in the above recall that @xmath56 is the average quality of representation @xmath11 for video @xmath7 .
thus , the objective expresses the video quality delivered to the complete hcn .
in the second term we have the quality for the users associated to the mbs since they can not transmit during an abs .
for the first set of constraints we have to recall that the fraction of the blank abs resources available for the pbss ( there is resource re - use across the pbss ) is @xmath0 .
this leads to : @xmath57 in the above we excluded again the mbs since it can not transmit during an abs . during the rs all the bss
transmit : @xmath58 when a particular representation @xmath11 is selected , the average rate @xmath59 in bits / sec that must be sustained by a user @xmath7 is less than the rate that can be achieved during both the abs and rs .
also the resources allocated during abs and rs will determine the average rate .
the above can be formally written as : @xmath60 in this section this constraint is based in , and in this form it ensures that the average number of rebuffering time over the complete video playback is zero . also
, accounts for the startup delay as specified in .
we will delve into the extension for dash in the next section .
we also have that resources can not be allocated to a video representation @xmath11 if it is not actually selected : @xmath61 we also need the integer constraints according to which only one video representation @xmath11 can be used for each user .
thus : @xmath62 during the regular slots the pbss can also transmit together with the mbs , albeit with lower spectral efficiency . in this case the rate will be lower .
however , we must ensure that across abs and rs the same video representation is used : @xmath63 the last condition comes from the observation that it is not practical to transmit one video quality during abs and a different during the rs , since these two types of resources alternate at the phy in the order of milliseconds @xcite .
this problem formulation clearly constitutes a non - convex mixed integer linear program ( milp ) .
hence , it is np - hard while it does not map to a well - known structure that can be solved with fast pseudo - polynomial algorithms ( e.g. , knapsack forms ) .
this last aspect can also be seen fairly easily since we have that @xmath0 , which is the capacity of the knapsack in , , is also an optimization variable .
the second issue is the evident need for distributed computation .
these aspects make the problem computationally challenging . despite this difficult challenge at first sight
, we notice that at this stage of our work we are interested to calculate the average rate allocation and tdrp during the complete streaming session .
this means that in practice the final algorithm does not have to provide a solution in the order of seconds , but minutes .
for this reason , instead of designing heuristics , we resort to a solution approach with a primal - dual approximation algorithm that converges asymptotically to the optimal solution @xcite . to solve this problem we apply first primal decomposition on @xmath0 .
primal decomposition consists of setting a constant value to the coupling variable @xcite . for a constant value for @xmath0
we notice that the original problem is decomposed into a master problem @xmath64 , and several problems denoted as @xmath65 ( each one for each bs @xmath2 ) .
we use lagrangian relaxation to solve @xmath65 .
the lagrangian after relaxing the coupling constraints for @xmath65 is : @xmath66 in the above @xmath67 are constraint vectors , , constraint is expressed with vector @xmath68 , and , are written as the constraint vectors @xmath69,@xmath70 , while @xmath71,@xmath72 correspond to , .
finally @xmath73 corresponds to .
the dual variables @xmath74 are all in row vector form in order to avoid the need for a transpose superscript .
now by using this relaxation , and by packing all the lagrangian multipliers for pbs @xmath2 as the single vector @xmath75 , the dual problem can be written as follows : @xmath76 a constant @xmath0 after primal decomposition has further implications .
in particular the problem in is decomposed into subproblems for the abs and rs
. we also notice from @xmath77 that these subproblems can be further decomposed for both the abs and the rs .
first we have a ra problem @xmath78 : @xmath79 also a vqs problem , denoted as @xmath80 : @xmath81 thus , we have two linear ra and two integer vqs problems that are solved by each bs as we explain next . the dual problem is solved in an iterative fashion , using a primal - dual lagrange method that can allow us to reach an asymptotically optimal solution @xcite .
the central concept of the primal - dual algorithm is to initialize first the dual variables to zero , and then to solve subproblems @xmath82 to obtain the currently optimal solution for iteration @xmath83 as @xmath84 , and @xmath85 . besides our problem - specific details described in this subsection , further details regarding the primal - dual method and its asymptotically optimal convergence property can be found in @xcite .
first we focus on subproblem that is linear program .
hence , it can be efficiently solved using standard convex optimization techniques @xcite .
the bs also solves the second subproblem , i.e. , the integer linear program ( ilp ) in for identifying the currently optimal video representation for each user .
this problem is solved in pseudo - polynomial time using dynamic programming ( dp ) @xcite .
its speed of convergence is evaluated collectively for the complete system in our performance evaluation , while the time complexity of the complete jtravqs is discussed in a later subsection . after solving the subproblems , and given the current result @xmath86
, we employ a sub - gradient method @xcite to update the dual variables .
a representative calculation is presented for the dual variables of the very first constraint : @xmath87^+ \label{eq : dual_update2}\ ] ] in the above , the term in the parenthesis is the sub - gradient , @xmath88^+$ ] denotes the projection onto the non - negative orthant , and @xmath89 is the step size at iteration @xmath83 .
similarly we define the update rules and the subgradients for the remaining dual variables . in each iteration @xmath83 , the dual objective is improved using the subgradient update and accordingly the primal relaxed problems @xmath82 are solved again in order to update the primal variables ( which are then used in the subsequent dual objective update ) . for the final step ,
the dual variables for constraints of each @xmath65 are provided to the master problem @xmath64 that has to be solved now for the optimal @xmath0 at the central controller ( cc ) of the system .
recall that for the master problem we applied primal decomposition .
it is thus solved very efficiently by collecting only the resource prices for @xmath0 , i.e @xmath90 , from each pbs in order to form the global subgradient @xcite . in practice
we only transmit the local subgradient @xmath91 from each bs .
the current value of @xmath0 is updated as follows @xcite : @xmath92 } _
\text{global subgradient}\big ] ^+ \label{eq : primal_update}\ ] ] now @xmath93 is a vector that can be selected as before in order to control the speed of convergence @xcite .
the time complexity of the discrete knapsack problem denoted by @xmath80 , when solved with dynamic programming , is polynomial with respect to the size of the problem @xmath94 for each bs @xmath2 , but for bounded knapsack capacity .
hence , in our case it is @xmath95 , i.e. , linear .
this is because from we notice that the capacity of the knapsack is 1 , in other words only one item ( video representation ) can be selected .
@xmath78 is a lp and so polynomial with respect to the number of associated users @xmath96 , i.e. , @xmath78 is @xmath97 .
hence , the time complexity of @xmath80 may be better than that of @xmath78 .
we conclude that the worst execution time of one iteration of the jtravqs algorithm , as a function of its inputs , can be expressed as : notation have to be multipled with the execution time of the fundamental algorithm operation . ]
@xmath98 where @xmath99 is the delay between bs @xmath2 and the cc , @xmath100 corresponds to the execution time of the primal update ( one vector multiplication and one addition ) and @xmath101 is the delay for communicating the new primal update to the bss . in our simulation
we also considered that the backhaul links have the _ worst case _ transmission delay of @xmath99=60ms @xcite .
for 100 iterations the total delay until the optimal solution is reached is 6 seconds which means that the calculations have to take place within 4 seconds in order to reach a solution within a 10 second period .
this is well within the capabilities of modern processors for solving the discussed lp and ilp . also , each bs @xmath2 communicates only @xmath102 to the cc which constitutes a negligible overhead .
the good performance of the algorithm , allow us to investigate its use in shorter time scales next .
* motivation . * in a network it is possible that channel conditions and users are more dynamic . in this case
the bitrate of the transmitted video should be adapted .
one way to accomplish that is dash . with dash a video
is stored as a sequence of short duration ( typically 2 - 10 sec ) video segments @xcite .
each segment may be available at different sizes , snr qualities , spatial resolutions , frame rates .
however , it has been shown that allowing the client to be fully responsible for requesting video segments ( a pull - based system ) , after estimating the variations in the end - to - end throughput , results in significant waste of resources @xcite . in this paper ,
our optimization framework at the bs is responsible for the choice of the optimal video representation .
this is also a realistic option since dash does not specify where video adaptation occurs .
* enhanced system model . *
we define the term _ slot _ as the period that the problem is solved and its decisions are enforced ( see fig .
[ fig : enforcement - of - ra ] ) . without loosing generality
we assume that jtravqs - dash is solved during a slot with a duration of 10 seconds .
since the problem is solved for every slot , the instance of the problem currently solved is also indexed by @xmath31 .
the result is that the jtravqs - dash problem is solved during slot @xmath31 to calculate the optimal ra and vqs for slot @xmath103 .
the algorithm requires several iterations as before , that are indexed also by @xmath83 . regarding the input parameters to jtravqs - dash @xmath104
is our estimate of the tcp throughput of user @xmath7 during slot @xmath31 according to @xcite , and @xmath105 the set of associated users .
this approach for modeling @xmath104 , is consistent with the behavior of dash that uses tcp for downloading segments .
hence , contrary to related work our approach is more realistic with respect to @xmath104 @xcite . the video quality model in
is used with @xmath34 denoting the quality of the remaining segments that should be transmitted .
let us finally define some minimal additional notation since the optimization variables must be indexed by slot @xmath31 : @xmath106 and @xmath107 .
now @xmath108 indicates that in slot @xmath31 segments from representation @xmath11 will be transmitted to user @xmath7 .
the same algorithm is used for solving jtravqs - dash but the two subproblems are adapted . *
dash rate allocation ( dashra ) problem . * the most important aspect is the re - formulation of constraint that is now indexed by the slot @xmath31 , and is packed into constraint vector @xmath109 : @xmath110 re - writing for the dash case yields the problem @xmath111 : @xmath112 in @xmath113 is the average bitrate that must be sustained by the remaining segments of the @xmath11-th representation similarly to , .
the key difference from the initial problem is parameter @xmath114 .
this is the total duration of the playable video in seconds that user @xmath7 has in its buffer at the start of slot @xmath31 and is denoted as @xmath115 , minus the playable video that the slowest user has in its buffer , i.e. , @xmath116 .
each user updates the estimate of the playable video as : @xmath117 .
this parameter ensures that users that have received lower volume of data are effectively prioritized .
hence , if the ra decision for user @xmath7 in the @xmath118-th slot is @xmath119 , then at the start of slot @xmath118 the bs can calculate @xmath115 , since it knows the result of ra and of course the duration of video that will be played . to summarize , this is effectively a rebuffering constraint that contains the differential buffer information .
* dash client model example . *
to explain the playback model for the dash client and the setting of @xmath114 , let us use a specific example with clients that have different playback buffer contents as illustrated in fig .
[ fig : enforcement - of - ra ] .
the number of transmitted segments depend on the value of @xmath119 .
also in this example we consider the downloading of complete segments for exposition purposes but our model supports partially downloaded segments . for
user 1 assume that it is the client that is lagging behind from the rest and the playble video it has in its buffer at the start of slot @xmath118 is @xmath120=0 .
hence , at the start of slot @xmath118 it will rebuffer until it receives the segment and after it finishes , the video player enters the playback mode . also @xmath121=@xmath120 + 10 - 10=0 , and @xmath122=@xmath123=0 . at the start of slot @xmath118 user 2
has 20 seconds worth of video , while during @xmath118 it will receive two additional segments leading to @xmath124=20 + 20 - 10=30 , and @xmath125=@xmath126=30 .
hence , by having allocated more resources with @xmath127 the result is a pre - fetching of data . for user 3
similarly we obtain @xmath128=10 + 10 - 10=10 and @xmath129=10 .
* dash video quality selection ( dashvqs ) problem .
* now the vqs problem is solved by adding one constraint in the problem .
we reformulate the @xmath130 problem to @xmath131 that is also solved over the @xmath118 slot : @xmath132 note that for partial downloading , if in the next slot dashvqs identifies that a lower or higher video quality is transmitted , then pre - buffered data are not discarded but the unfinished segment is received . the new decision for the video quality
is enforced when a new segment will be transmitted .
the dual variables are updated as before and the sub - gradients are similarly modified based on the new constraint .
in this section , we present a comprehensive evaluation of the proposed algorithms comprising our framework through custom matlab simulation . our simulator performs a precise phy - level simulation of wireless packet transmissions . * jtravqs evaluation . *
the parameter settings for our simulations are set as follows .
downlink mbs and pbs transmit power are equal to 46dbm and 30dbm respectively @xcite .
distance - dependent path loss is given by @xmath133 , where @xmath134 is the distance between two nodes in km @xcite , and the shadowing standard deviation is 8 db .
the user speed is 3 kmph ( quasi - static as we already stated ) , and average cqi is provided every 10 minutes . the macrocell area is set to be a circle with radius equal to 1 km .
the wireless channel parameters include a channel bandwidth of @xmath135=20 mhz , noise power spectral density of @xmath20=@xmath136 watt / hz , while the same rayleigh fading model was used for all the channels .
packets of 1500 bytes are transmitted at the phy , while the optimal mcs is calculated according to .
the user distribution and picocell locations are random and uniform within the macrocell .
we set the biasing threshold to 0 db for all the systems to calculate @xmath4 .
the user population increases up to a number of 200 to evaluate the performance in networks that continuously become more dense , consistently with the recent trends @xcite . for the pfra system we configured the users to request randomly and uniformly one of the available video representations ,
while for jtravqs users request randomly and uniformly one of the available videos .
the video content used in the experiments consists of 26 cif ( 352x288 ) , and high definition ( 1920x1080 ) sequences that were encoded with svc h.264 as a single layers @xcite .
the videos are compressed at 30 fps and different rates ranging from 128 kbps and reaching values@xmath1377 mbps .
the frame - type patterns were g16b1,g16b3,g16b7,g16b15 , i.e. , there are different numbers of b frames between every two p frames and a gop size is always equal to 16 frames . regarding the presentation of the results , fig . [
fig : results2 ] shows the average video quality ( in terms of the representation @xmath11 ) that is delivered to the picocell and macrocell users .
for example one data point that has the value 3.2 indicates that on average the users received the quality representation 3.2 .
hence , higher values indicate that the users received on average higher video quality representations .
the data points in these figures correspond to different values of @xmath0 . also , the data points correspond to the average ( mean ) of all the measurements for 100 randomly generated topologies .
the sample variance for this set the measurements is between 0.1 and 0.2 which is fairly small compared to the value of the mean and its difference between all the tested systems .
* video quality . * for this set of results we present the average video quality of the picocell users versus the average video quality of the users associated to the macrocell ( only from the mbs to its associated users ) for different constant values of @xmath0 to illustrate the impact of different tdrp .
the results for all systems can be seen in fig .
[ fig : results2](a , b ) for @xmath138=0.5 .
jtravqs is superior when compared to pfra for high user density and low pbs density in fig .
[ fig : results2](a ) . as the number of the pbss is increased to 8 in fig .
[ fig : results2](b ) , all the systems can achieve higher performance .
the reason is that fewer users are associated to each picocell and so a higher communication rate is available for each user under any scheme .
so more picocells leads to better results due to the higher available rate per user as expected .
another important result is that for constant pbs density ( either 4 or 8) , we have higher gain as the user population grows .
the reason is the higher importance of optimal rate allocation , since the rate of a single pbs is shared among several users .
for example in the very left data point of fig .
[ fig : results2](b ) performance improvement of jtravqs over pfra is 21% for 100 users and 36% for 200 users ) .
of course the average video quality is reduced for all systems since more users are present . also note that in the left part of the @xmath139 axis , where all the resources are practically allocated to the picocells ( @xmath140 ) , we observe the maximum possible video quality in the network . in this regime , the performance gap between jtravqs and the other systems is increased as the number of picocells and users is increased .
another important result in the same figure , is related to the optimal @xmath141 .
it is indicated with a dashed line that is intersected with representative performance curves .
this shows that the interpretation of the optimal tdrp with jtravqs , that is denoted as @xmath141(jtravqs ) , results in higher value for @xmath141 when compared to @xmath141(pfra ) by 22% ( highlighted with the horizontal arrow ) .
also if we assume that the system executes first pfra to calculate the optimal tdrp indicated as @xmath141(pfra ) , and then perform ravqs @xcite with this fixed value , the result is an average quality equal to 4.3 ( the gain is highlighted with the lower vertical arrow ) . however , our complete system calculates the optimal operating point indicated with @xmath141(jtravqs ) in the figure which gives an average quality equal to 5.1 , a performance difference of 18.6% ( the gain is highlighted with the upper vertical arrow ) .
our system offers significant performance increase for @xmath138=0.5 but the benefits are more important when @xmath138=0.75 in fig .
[ fig : results2](c , d ) .
note that for @xmath138=0.75 the slope of the curve is reduced as the @xmath0 is decreased .
the benefit is because we have a higher number of users that can be optimized under jtravqs .
also in this case the benefits are even more important when the fraction of the resources @xmath1 that the macrocell uses is below 50% ( left part of the @xmath139 axis ) since this gives more resources to the highly spectral efficient links in the picocells to be used and so a higher communication rate is possible . *
fraction of optimized users .
* now an interesting set of results is obtained for different values of @xmath138 .
we notice in fig .
[ fig : results6](a ) that as this fraction is reduced , jtravqs essentially degenerates to the pfra system .
nevertheless , we still obtain significant benefits even for ratios of @xmath138 around 30% since few users are enough for jtravqs to be able to improve the overall system performance . * primal - dual convergence . * the convergence speed of jtravqs versus the number of iterations
is illustrated in fig .
[ fig : results6](b ) .
results for the algorithm execution are shown for a specific fixed number of picocells and user population .
the results for the primal - dual algorithm used for jtravqs show that the convergence is achieved within 150 iterations . also fig .
[ fig : results6](b ) illustrates another important aspect of our system : if the system uses fewer discrete quality representations for each video file , it allows the faster convergence of the algorithm . * jtravqs - dash evaluation .
* the performance of jtravqs - dash is evaluated with the same setup as before , that is however augmented when necessary .
now we present results for the playback performance : the time that a client is rebuffering in seconds , and the number of rebuffering events per minute . to ensure fairness
, we calculate first the optimal solution with jtravqs .
then , the minimum video quality for the jtravqs - dash system is set equal to jtravqs .
this ensures that jtravqs - dash delivers at least the same video quality and the question is then to evaluate its ability to minimize rebuffering .
for static user conditions the results are illustrated in fig .
[ fig : results6](c ) .
we draw a notched box plot of the rebuffering time for all the clients in the upper part of fig . [
fig : results6](c ) , and the number of rebuffering events in the lower part of fig .
[ fig : results6](c ) .
the notch here marks the 95% confidence interval for the _ median_.
all the systems perform well for 20 and 40 users since high capacity is available in the network .
however , for higher user densities the buffering time with the baseline jtravqs is increased by a factor that is worse than linear .
the same is true for pfra that is slightly worse . with a higher number of 8 pbss , rebuffering time
is improved for jtravqs because of higher available capacity .
the same is true for the number of rebuffering events / minute that is a very high number for the first three systems we discussed ( 2 - 3 events in a 10 minute period ) .
hence , the higher capacity in the network achieved with 8 pbss , simply delays the inevitable sharp increase but only of the rebuffering time .
this result has an interesting interpretation for mnos : with increasing user density , expanding the network with more small cells improves marginally the rebuffering time and practically not at all the rebuffering events .
this is in contrast to the video quality that can achieve significant improvement in our earlier plots . for extra gains , solutions with buffer - awareness
are required .
better results are obtained for ravqs - dash again with a constant @xmath0=50% .
this is effectively a system configuration that encompasses the main features of the dash - aware streaming literature for single cell networks , e.g. @xcite , where the rate allocation and video quality are optimized by considering the buffer contents , but the overall communication resources are constant .
recall than @xmath0=50% is the optimal @xmath0 under a pfra metric for a population of 100 users and 4 pbss ( as illustrated in our earlier figures ) .
buffer - awareness can indeed reduce the rebuffering when compared to the previous systems , while it can also reduce the variations of playback buffering for the users ( we have more predictable performance ) .
the proposed jtravqs - dash system illustrates that it can reduce the time spent in rebuffering by over 50% when compared to the previous system , while the number of rebuffering events is reduced even more significantly ( 1 event/25 min . vs. 1 event/10 min . ) .
hence , for a hcn the fixed tdrp is not the best option even if we design a dash - aware system .
also , our overall results indicate that using a fixed tdrp has worse consequences in the rebuffering time / events of dash than on the video quality ( e.g. , the results illustrated in fig .
[ fig : results2 ] ) . finally , we evaluate our system for a time - varying user population based on results from a real 3 g network reported in @xcite .
we simulate an 8 hour period between 4 pm and 12 am , with a user peak occurring around 9 pm @xcite . during this peak ,
the number of users is nearly 30% higher than the number of users at 4 pm and 12 am .
hence , in our simulation we set accordingly @xmath142 ( the increase and decrease are approximated as linear in time as shown in @xcite ) . in the results in fig .
[ fig : results6](d ) we present in the @xmath139 axis the peak number of users .
@xmath104 for each user @xmath7 is also affected since tcp shares equally the communication rate among the competing traffic .
also , we only compare different flavors of jtravqs - dash since the previous systems are not designed for a dynamic network .
first , we observe again that the systems with fixed resource partitioning @xmath0=50% have worse performance .
second , we notice that the rebuffering time is higher when the fraction of optimized users is also high and equal to @xmath138=0.75 .
this means that with increasing user density in a network with user churn , optimizing a lower fraction @xmath138 of the users increases the dash playback quality of the optimized users by a significant amount for jtravqs - dash .
this is an important result since it provides a tool for an mno to differentiate qoe in terms of rebuffering to various users .
in this paper , we presented a framework for improving the quality of video streaming in a hcn that employs tdrp .
tdrp is essential for the efficient operation of hcns and when high quality video distribution enters the game , efficiency becomes even more important .
our framework addressed precisely this challenge , i.e. , it ensures optimal and video - aware allocation of resources in hcns that apply tdrp .
we formulated this problem in a linear non - convex formulation for which we proposed a primal - dual approximation algorithm .
our problem was decomposed into several problems that included a convex rate allocation problem , and a binary ilp for optimal video quality selection .
an extensive performance evaluation under different hcn configurations highlighted the value of our framework for obtaining video quality improvements .
another implication of our solution approach is that it can be solved very fast .
this allowed us to augment it to support the more challenging case of dash . in this case
significant additional improvement in terms playback performance was obtained .
s. deb , p. monogioudis , j. miernik , and j. seymour , `` algorithms for enhanced inter - cell interference coordination ( eicic ) in lte hetnets , '' _ networking , ieee / acm transactions on _ , vol .
22 , no . 1 ,
137150 , feb 2014 .
d. kosmanos , a. argyriou , y. liu , l. tassiulas , and s. ci , `` a cooperative protocol for video streaming in dense small cell wireless relay networks , '' _ signal processing : image communication _ , vol
. 31 , pp . 151160 , february 2015 .
f. dobrian , v. sekar , a. awan , i. stoica , d. joseph , a. ganjam , j. zhan , and h. zhang , `` understanding the impact of video quality on user engagement , '' in _ proceedings of the acm sigcomm conference _ , 2011 .
w. wu , a. arefin , r. rivas , k. nahrstedt , r. sheppard , and z. yang , `` quality of experience in distributed interactive multimedia environments : toward a theoretical framework , '' in _
acm multimedia _ , 2009 .
j. chakareski , j. apostolopoulos , s. wee , w. tian tan , and b. girod , `` rate - distortion hint tracks for adaptive video streaming , '' _ circuits and systems for video technology , ieee transactions on _ , vol . 15 , no . 10 , pp .
12571269 , oct 2005 . | heterogenous cellular networks ( hcn ) introduce small cells within the transmission range of a macrocell . for the efficient operation of hcns
it is essential that the high power macrocell shuts off its transmissions for an appropriate amount of time in order for the low power small cells to transmit .
this is a mechanism that allows time - domain resource partitioning ( tdrp ) and is critical to be optimized for maximizing the throughput of the complete hcn . in this paper
, we investigate video communication in hcns when tdrp is employed . after defining a detailed system model for video streaming in such a hcn
, we consider the problem of maximizing the experienced video quality at all the users , by jointly optimizing the tdrp for the hcn , the rate allocated to each specific user , and the selected video quality transmitted to a user .
the np - hard problem is solved with a primal - dual approximation algorithm that decomposes the problem into simpler subproblems , making them amenable to fast well - known solution algorithms .
consequently , the calculated solution can be enforced in the time scale of real - life video streaming sessions .
this last observation motivates the enhancement of the proposed framework to support video delivery with dynamic adaptive streaming over http ( dash ) .
our extensive simulation results demonstrate clearly the need for our holistic approach for improving the video quality and playback performance of the video streaming users in hcns .
heterogeneous cellular networks , small cells , intra - cell interference , video streaming , video distribution , dash , rate allocation , resource allocation , optimization , 5 g wireless networks . |
in color - magnitude diagrams ( cmds ) of open clusters , stars are observed between the blue straggler stars ( bsss ) and the red giant branch ( rgb ) .
we call these yellow straggler stars ( ysss ) . in m67 ,
a @xmath44 gyr , solar - metallicity open cluster @xcite , there are four 3d kinematic members found in this cmd region .
three are binary systems , and one is a single star @xcite . while some ysss might be explained as the combined light of two cluster stars ( e.g. , a red giant - blue straggler binary ) , many may be evolved bsss @xcite
. these stars would be post - main - sequence stars more massive than the main - sequence turnoff ( msto ) .
bsss are thought to form from mass transfer in binary systems @xcite , stellar collisions during dynamical encounters @xcite , or binary mergers induced by kozai cycles @xcite . in m67 ,
one of the four ysss has a helium white dwarf ( wd ) companion , suggesting it is an evolved bss formed from mass transfer @xcite .
asteroseismic analysis of _ kepler _ stars have uncovered red giants in open clusters more massive than the cluster standard @xcite . these stars are called evolved blue straggler stars ( e - bsss ) .
they fall in or near their clusters red clumps , but ysss may be bluer examples of the e - bsss , making them more identifiable in cluster cmds . mass measurements for ysss are needed to confirm this hypothesis .
recent data from the _ kepler _ k2 mission allow for the asteroseismic measurement of stellar masses and radii in m67 @xcite , providing an exciting opportunity to study the ysss in this cluster . here
we report the first determination of an asteroseismic mass and radius for the primary star in the m67 binary yss s1237 @xcite .
we summarize radial - velocity ( rv ) and x - ray observations of s1237 , and use the spectral energy distribution ( sed ) to constrain the mass of the secondary star .
we compare the asteroseismic mass to the primary mass implied from its cmd position , and comment on possible formation pathways for this star .
m67 was observed during campaign 5 of the _ kepler _ k2 mission , providing high - precision 75-day light curves for targets in the cluster .
light curves were obtained using the method of @xcite .
the light curve of yss s1237 shows solar - like oscillations and our seismic analysis followed the same approach as @xcite .
we processed the light curve of s1237 as in @xcite and used the pipeline of @xcite to extract the large frequency separation ( @xmath5 ) ) and frequency of maximum power ( @xmath6 ) from the fourier frequency spectra of the light curve . from these global frequencies
, we used the asteroseismic scaling relations , @xmath5 @xmath7 and @xmath6 @xmath8 ) ( e.g. @xcite ) , to determine a mass and radius for the star . . *
( b ) * folded and smoothed spectrum of the central @xmath9 range around @xmath10 ( black curve ) .
the red dashed line shows an empirical model representing the average of a large ensemble of red giants used to measure @xmath11 and determine the location of the different modes @xcite .
the arrow indicates the offset from zero of the radial modes .
[ ps],title="fig : " ] . *
( b ) * folded and smoothed spectrum of the central @xmath9 range around @xmath10 ( black curve ) .
the red dashed line shows an empirical model representing the average of a large ensemble of red giants used to measure @xmath11 and determine the location of the different modes @xcite .
the arrow indicates the offset from zero of the radial modes .
[ ps],title="fig : " ] lll + ra & 8:51:50.20 & + dec . & 11:46:07.0 & + epic i d & 211408357 & + wocs i d & 1015 & + sanders i d & s1237 & + + period ( p ) ( days ) & @xmath12 & @xmath13 + eccentricity ( e ) & 0.087 & @xmath14 + amplitude ( k ) ( km s@xmath15 ) & 5.024 & @xmath16 + f(m ) ( m@xmath1 ) & 0.0091 & @xmath17 + + period ( p ) ( days ) & @xmath18 & @xmath19 + eccentricity ( e ) & 0.105 & @xmath14 + amplitude ( k ) ( km s@xmath15 ) & 5.03 & @xmath20 + f(m ) ) ( m@xmath1 ) & 0.0091 & @xmath17 + + @xmath10 ( @xmath21hz ) & 114.89 & @xmath22 + @xmath5 ( @xmath21hz ) & 8.32 & @xmath23 + radius ( r@xmath1 ) & 9.11 & @xmath24 + mass ( m@xmath1 ) & 2.87 & @xmath25 + surface gravity ( log @xmath26 ) & 2.98 & @xmath27 0.01 + + @xmath28 & 1.0057 & + radius ( r@xmath1 ) & 9.21 & @xmath27 0.19 + mass ( m@xmath1 ) & 2.94 & @xmath27 0.24 + + @xmath28 & 1.0089 & + radius ( r@xmath1 ) & 9.27 & @xmath27 0.19 + mass ( m@xmath1 ) & 2.97 & @xmath27 0.24 + @xmath29 @xmath30 the @xmath5 scaling relation needs a small metallicity- and temperature - dependent correction ( e.g. @xcite ) . to derive this correction factor , @xmath28
, we adopt @xmath31 k and [ fe / h]= 0.072 @xmath27 0.033 ( from apogee ; @xcite ) , which we feed into the correction interpolator of @xcite .
we derive two correction factors , one assuming the star is helium - core burning ( heb ) and the other assuming it is an rgb star .
we obtain r = 9.27 @xmath27 0.19 r@xmath1 and m = 2.97 @xmath27 0.24 m@xmath1 for a heb star , or r = 9.21 @xmath27 0.19 r@xmath1 and m = 2.94 @xmath27 0.24 m@xmath1 for an rgb star ( table [ asteroparams ] ) .
the errors on mass and radius come from propagating errors on @xmath32 , @xmath5 , and @xmath6 through equations 1 and 2 .
our results seem to indicate this is a heb star .
a 2.6 - 2.9 m@xmath1 star with this radius would be in the hertzsprung gap , and thus is unlikely to be observed .
the power spectrum also shows broader peaks and more blended @xmath33 and @xmath34 modes than typical of rgb stars .
we show the full power spectrum in figure [ ps]a , as well as the central @xmath35 around @xmath10 folded on @xmath5 and smoothed with a gaussian as in @xcite ( figure [ ps]b ) .
the folded spectrum shows a broad series of multiple peaks around the location of the dipole ( @xmath36 ) mode .
this is indicative of a large period spacing between dipole mixed modes @xcite and of a larger coupling between the envelope and core of the star @xcite , both suggestive of a heb star . in figure
[ ps]b we also measure @xmath11 , the offset from zero of the radial modes ( @xmath37 ) , by correlating the spectrum with a model profile following @xcite . we find @xmath38 , outside the range for an rgb star ( @xcite , fig .
updating the orbital solution from @xcite with additional rvs from the wiyn open cluster study ( wocs ; @xcite ) , we find the orbital parameters do not change significantly ( see table 1 ) .
the system is a near - circular 698-day period spectroscopic binary .
the system is single - lined in the wocs spectra ( @xmath39 ) . given the small amplitude of the orbit and the fact that any ms companion would be more than 2 magnitudes fainter than a @xmath40 r@xmath1 giant primary
, we do not expect to see the features of the secondary at the signal to noise of the wocs spectra , .
higher resolution , high signal - to - noise spectra may allow detection of the secondary . for a primary of 2.9 m@xmath1 , the binary mass function ( @xmath41 )
indicates a secondary with @xmath42 m@xmath1 .
s1237 has a 0.3 - 7 kev luminosity of l@xmath43= 5.5 @xmath44 ergs s@xmath15 , comparable to what is expected from rapid rotation @xcite .
however , wocs spectra of the primary provide an upper limit on the @xmath45 sin @xmath46 of 10 km s@xmath15 set by the instrumental broadening and the binary is too wide for rapid rotation to be expected from tidal synchronization . @xcite also detect s1237 as an x - ray source and find no satisfactory explanation .
we discuss a possible explanation for this x - ray emission in section 6 .
we have assembled an sed from existing ubvi photometry @xcite , the two - micron all sky survey ( 2mass ; @xcite ) , wide field infrared explorer ( wise ; @xcite ) , and the galaxy evolution explorer ( galex ; @xcite ) .
we use @xmath47 minimization to fit the observed sed to a grid of models .
we first fit a model of the primary to the 2mass and wise data .
these ir measurements should be dominated by the flux from the primary , whereas the optical and uv flux may be contaminated by flux from a ms secondary .
we model the primary with a 4500 - 5500 k giant spectrum @xcite in steps of 50 k in order to bracket the the apogee @xmath32 of @xmath48 k. we use a grid of radii between 8.0 - 10.0 r@xmath1 in 0.1 r@xmath1 increments .
we fix the surface gravity to log g= 3.0 from asteroseismic scaling ( table 1 ) .
we use distances of either 800 , 850 , or 900 pc based on the range given in the literature @xcite .
we adopt e(b - v)= 0.041 @xcite and the extinction curve of @xcite .
the best - fit sed temperature and radius only fall within the 3@xmath49 errors on the apogee temperature and the asteroseismic radius if we use a distance of 900 pc . using 900 pc ,
we find a best - fit of 8.8 r@xmath1and 4900 k ( see figure 2b ) .
using this model for our primary , we find a significant flux excess over the model in all of the bluer bands ( fuv , nuv , u , b , v ) indicating the presence of a secondary ( see figure 2a ) . fixing the primary to these best - fit values ,
we model the secondary using a fixed surface gravity of log g= 4.5 ( typical for a ms star ) , a grid of temperatures between 5000 - 8000 k in steps of 250 k , and secondary radii between 0.5 r@xmath1 - 2.0 r@xmath1 in steps of 0.1 r@xmath1 .
we find that a variety of secondary models result in @xmath47 values near the minimum with temperatures of 6250 - 8000 k and radii of 0.5 - 1.8 r@xmath1 , but if we restrict the radius to fall between the zams radius and the radius at 4 gyr for a given temperature , we find the best fit to be a @xmath46250 - 6750 k , 1.2 - 1.6 r@xmath1 star ( figure 2c , d ) .
this range includes both stars on the upper main - sequence in m67 , as well as blue straggler stars near the msto . for simplicity
, we assume the star is on the upper ms when considering formation scenarios , but the possibility that the secondary is a blue straggler is worth follow up , i.e. with higher resolution spectra that could detect the secondary .
we note there is a galex fuv excess in the sed ( figure 2c ) , but we do not consider this excess to be meaningful .
castelli - kurucz models have poor resolution in the fuv , and a few unresolved emission lines can change the flux substantially .
in fact , ms stars near the turn off in m67 have widely varying galex fuv fluxes , and several are observed to have comparable fuv magnitudes .
in figure [ cmd ] we decompose the light of the binary to show the location of the primary and the secondary in a cmd .
we show two possible deconvolutions : an ms secondary and a bss secondary .
the secondary is fixed to either a 6750 k bss ( red circle ) , or a 6250 k ms star near the turnoff ( orange circle ) .
the primary is a @xmath50 k giant .
we compare the cmd position of the primary and secondary to evolutionary tracks using modules for experiments in stellar astrophysics ( mesa ; @xcite ) .
we use the test suite case ` 1m_pre_ms_to_wd ` changing only the mass and turning off rgb and agb wind mass loss .
the cmd location of the primary is fainter than expected from mesa models for a giant with a mass and radius within @xmath51 of the asteroseismic measurements . to bring the photometry and asteroseismology into agreement requires using a distance at the highest range of those determined in the literature for m67 ( 900 pc ) , and a primary mass and radius of @xmath52 m@xmath1 and @xmath53 r@xmath1 , more than @xmath3 below the asteroseismic values . the bss and msto secondary fall on evolutionary tracks for a @xmath54 and @xmath55 m@xmath1 star , respectively .
while using a distance of 900 pc to m67 creates the best agreement between the asteroseismic mass and the photometry for s1237 , @xcite find an asteroseismic distance to m67 of @xmath56 pc based on the entire sample of red giants in m67 . using 816 pc results in a photometric mass and radius for the primary more than @xmath57 below the asteroseismic values .
depending on the distance used , the asteroseismic results may be @xmath58 and @xmath59 higher than the photometric mass and radius , respectively .
some recent studies comparing _ kepler _ masses and radii from uncorrected asteroseismic scaling relations and rgb eclipsing binaries have found systematic differences between the two measurements ( e.g. @xcite , @xcite ) similar to the size of the discrepancy we find .
we have used corrected scaling relations , which should provide better agreement @xcite , and @xcite find that the average rgb mass in m67 from corrected scaling relations is consistent with eclipsing binary results for the cluster .
however , the scaling relations have not been well validated by independent measurements ( e.g. comparison to eclipsing binaries or cluster isochrones ) for heb stars at masses as large as 2.9 m@xmath1 , and it is therefore possible that the corrections that work for lower mass stars can not be extrapolated to stars of such high mass .
this uncertainty is not reflected in our errors .
we encourage more tests of the scaling relations for massive heb stars to shed light on this possible discrepancy .
alternatively , s1237 may be following a non - standard evolutionary track due to past mass transfer , a merger , or a stellar collision .
however , models of the evolution of mass - transfer products predict they return to close - to - normal evolution soon after mass transfer has ceased @xcite
. models of stellar collision products are bluer and brighter than standard evolutionary tracks due to enhanced mixing enriching the star in helium , although this difference is minor on the giant branch @xcite .
our results indicate the opposite : the star is less luminous and redder than expected given its asteroseismic mass .
apogee c , n , and fe abundances do not appear unusual , but a more careful abundance analysis may prove interesting .
using mesa we estimate the primary s age to be 450 myr if it is a 2.9 m@xmath1 heb star .
this is likely an overestimate , as it assumes the primary begins with a core that has not burned through any hydrogen . in reality , this star likely formed by adding material to a lower mass ms star to create a bss .
the ms progenitor would already have gone through @xmath43.5 gyr of evolution , perhaps resulting in a core substantially depleted of hydrogen .
this translates to a shorter lifetime for the star once it becomes a bss .
if we consider a 2.4 m@xmath1 star , @xmath42 @xmath49 lower than the asteroseismic value but consistent with the mass implied from its cmd position , the lifetime is @xmath4750 myr .
this longer lifetime would result in a more massive msto at the time of formation , and the lower mass makes it more plausible that the star formed from mass transfer or a merger .
we discuss this in greater detail in the following sections .
the m67 msto is 1.3 m@xmath1(@xmath60 m@xmath1 ; @xcite ; @xcite ) , and would have been only slightly more massive 450 myr ago .
mass transfer between two such stars would not create a star within the @xmath51 uncertainty on the mass of s1237 ( @xmath61 m@xmath1 ) .
if we consider a @xmath62 m@xmath1 primary the star could have formed from two 1.4 m@xmath1 stars ( the turnoff mass 750 myr ago ) and a mass transfer origin is worth considering .
the result of mass transfer from an agb star would be a blue / yellow straggler - wd binary , and thus some mass would be locked up in a wd companion . given the typical carbon - oxygen wd mass of 0.5 - 0.6 m@xmath1 , we would expect a primary with @xmath63 m@xmath1 , even under the unrealistic assumption of totally conservative mass transfer .
a lower mass helium wd is not possible because the orbital period of 698 days is too wide to be the result of rgb mass transfer @xcite .
additionally , the sed shows no clear evidence of a hot wd companion , though an older ( @xmath64 myr ) , cool wd would be compatible with the fuv photometry .
the system requires a ms companion to explain the sed , so including a wd would require the system to be a triple , e.g. a giant- wd binary with a wide msto tertiary .
the orbital solution and residuals show no indication of perturbation by such a third body . mass transfer leading to a merger between two ms stars could create a @xmath65 m@xmath1 star without a wd remnant .
such mergers may happen when magnetic braking shrinks the orbit of a close binary , but ms mergers may also result from angular momentum loss during a dynamical encounter or kozai - cycle oscillations ( see sections [ dynamical ] and [ kozai ] ) . [
cols="^,^,^,^,^,^,^,^,^,^,^,^ " , ] at least three turnoff mass stars are needed to make a 2.9 m@xmath1 primary , and another is required for the secondary . a binary - binary , single - triple , binary - triple , or triple - triple encounter could provide the four stars needed to create s1237 . if we consider a 2.4 m@xmath1 primary it is possible to form the primary from just 2 stars , making a single - binary encounter viable .
@xcite present a method to estimate bss production rates from stellar collisions .
we use their equations and parameters for m67 to calculate encounter rates and present them in table [ rates ] .
we also list the poisson probability ( @xmath66 ) that at least one encounter of each type would have occurred within the lifetime of s1237 , calculated using : @xmath67 here @xmath68 is the lifetime from section 5.1 , and @xmath69 is the encounter rate .
these rates suggest a few encounters can have occurred within the lifetime of s1237 , so @xcite indicate a collisional origin is possible .
we note that post - encounter binaries most commonly have higher predicted eccentricities than s1237 @xcite .
theory predicts s1237 should not have tidally circularized @xcite , so if s1237 was collisionally formed the low eccentricity is unusual .
one collision is sufficient to create a @xmath70 m@xmath1 primary , but to create a 2.9 m@xmath1 primary requires at least 2 collisions to occur during the encounter .
scattering experiments show that binary - binary encounters have significant probability ( @xmath71 ) of producing two or more collisions due to the increased cross - section for interaction after the adiabatic expansion of the initial collision product @xcite .
alternatively , s1237 could have undergone multiple collisional encounters .
n - body simulations of m67 have produced multiple ` super - bsss ' with masses more than twice the turnoff mass after back - to - back dynamical encounters @xcite . while the rates suggest such massive bsss should be rare , we know they are produced in open clusters .
in addition to s1237 , m67 hosts bsss s977
f81 ) and s1082 @xcite .
s977 has a cmd location indicative of a star more than twice the turnoff mass , and s1082 is a triple bss system with a combined dynamically - measured mass of @xmath72 m@xmath1 @xcite .
multiple collisions seems to be the most plausible mechanism to create these massive systems .
studies @xcite have proposed that kozai - cycle - induced mergers of an inner binary in a hierarchical triple may form bsss .
@xcite run a large set of monte carlo models to determine the effect of kozai cycles on the final orbital configuration of hierarchical triples .
they find the triples that lead to inner binary mergers resemble s1237 in several ways .
first , they find that mergers most often originate in triples with outer orbital periods ( @xmath73 ) of @xmath74 days , consistent with the 698-day period of s1237 .
second , they find a peak in the mass ratio ( @xmath75 ) between the tertiary and inner binary of 0.4 - 0.5 . assuming @xmath76
m@xmath1 and @xmath77 m@xmath1 , s1237 has a ratio of 0.45 .
third , merged systems with @xmath78 days have a fairly flat eccentricity distribution from 0.0 to 0.5 @xcite , consistent with the low orbital eccentricity of s1237 ( e= 0.087 ) .
the vast majority of these systems merge in less than 10 myr @xcite .
therefore , if s1237 was a primordial triple we would expect it to have undergone a kozai induced merger and evolved to a stellar remnant long ago .
however , an initially stable primordial triple may be perturbed by a passing star leading to later - in - life kozai oscillations .
alternatively , a hierarchical triple could have formed dynamically a few hundred myr ago , and quickly undergone a kozai induced merger to create s1237 . in either scenario ,
the maximum mass for the merger product would be twice the turnoff mass , or @xmath79 m@xmath1 , just outside the @xmath51 error on the asteroseismic mass . to produce
a 2.9 m@xmath1 star would require that one of the stars in the inner binary was already a bss .
we present asteroseismic measurements of a binary yss in m67 from _ kepler _ k2 data . using @xmath80-corrected scaling relations , we find the primary star is a 2.9 m@xmath1 , 9.2 r@xmath1 , helium - burning star .
this is more than twice the mass of the msto .
sed fitting determines the secondary is near the msto , either a normal ms star or a bss .
the sed and stellar models are only marginally consistent with asteroseismology , requiring a primary mass and radius 2 - 3 @xmath49 lower than the asteroseismic measurements .
this suggests that s1237 is either redder and less luminous than a standard for its mass , or the asteroseismic mass and radius are too large by at least @xmath81 and @xmath59 , respectively .
we encourage testing the asteroseismic scaling relations for massive helium burning stars for possible discrepancies . observed x - ray emission from this star
has not been explained .
given the masses of the primary and secondary @xmath82= 2.9 m@xmath1 , @xmath83=1.3 m@xmath1 ) , the mass function indicates the orbit is inclined at @xmath84 .
if the spin and orbital axes are aligned , the giant primary may be rapidly rotating without observable line broadening or seismic frequency splitting @xcite , thereby explaining the x - rays . alternatively
the x - ray emission may arise from a rapidly rotating secondary .
many x - ray sources in m67 with similar x - ray luminosities are rapid rotators located near the msto ( see fig . [ cmd ] ) .
most are tidally synchronized close binaries , which is not the case for s1237 , but rapid rotation is also expected after a recent stellar collision or mass transfer event .
we review possible formation models and suggest that a binary encounter leading to one or more collisions or mergers is a possible formation pathway .
collisions may have taken place during the dynamical encounter itself , or perhaps a kozai - induced merger of an inner binary in a dynamically formed triple occurred shortly afterward .
this is the first asteroseismic mass measurement for a yss and the second yss for which a formation pathway has been suggested .
s1040 , another yss in m67 , has a helium wd companion , suggesting it formed through mass transfer @xcite .
these two examples indicate that ysss are likely evolved bsss , and that like the bsss there are multiple mechanisms to create them .
this paper includes data collected by the kepler mission .
funding for the kepler mission is provided by the nasa science mission directorate . | yellow straggler stars ( ysss ) fall above the subgiant branch in optical color - magnitude diagrams , between the blue stragglers and the red giants .
ysss may represent a population of evolved blue stragglers , but none have the direct and precise mass and radius measurements needed to determine their evolutionary states and formation histories .
here we report the first asteroseismic mass and radius measurements of such a star , the yellow straggler s1237 in the open cluster m67 .
we apply asteroseismic scaling relations to a frequency analysis of the _ kepler _ k2 light curve and find a mass of @xmath0 m@xmath1 and a radius of @xmath2 r@xmath1 .
this is more than twice the mass of the main - sequence turnoff in m67 , suggesting s1237 is indeed an evolved blue straggler .
s1237 is the primary in a spectroscopic binary .
we update the binary orbital solution and use spectral energy distribution ( sed ) fitting to constrain the color - magnitude diagram ( cmd ) location of the secondary star .
we find that the secondary is likely an upper main - sequence star near the turnoff , but a slightly hotter blue straggler companion is also possible .
we then compare the asteroseismic mass of the primary to its mass from cmd fitting , finding the photometry implies a mass and radius more than @xmath3 below the asteroseismic measurement .
finally , we consider formation mechanisms for this star and suggest that s1237 may have formed from dynamical encounters resulting in stellar collisions or a binary merger . |
in classical geometry of numbers the elements of an ordinary continued fraction for a real number @xmath0 are obtained from a _ sail _ ( i.e. a broken line bounding the convex hull of all points with integer coefficients in certain cone ) . in present paper
we find broken lines generalizing sails to the case of continued fractions with arbitrary elements .
this in its turn leads to the definition of `` infinitesimal '' continued fractions , whose sails would be differentiable curves .
such `` infinitesimal '' continued fractions are defined by two density functions : areal and angular densities .
the areal density function has a remarkable physical meaning .
consider an observer at the origin and let the body move along the curve with the velocity inverse to the areal density function .
then the body moves according to the second kepler law with respect to the observer , i.e. with constant sector area velocity .
this paper is organized as follows . in the first section we study the classical case of ordinary continued fractions . in section 2
we expand the notion of the sail to the case of continued fraction with arbitrary elements .
further we show how to write continued fractions starting with broken lines .
we generalize the proposed construction of sails to the case of curves , give an analog of continued fractions , and show several examples in section 3 .
for a nice reference to general theory of continued fractions we suggest the book @xcite .
several works are devoted to geometry of continued fractions ( e.g. @xcite , @xcite ) and to their generalizations to multidimensional case ( @xcite , @xcite , @xcite , etc ) .
notice that the case of broken lines with integer edges discussed in @xcite is a particular subcase of geometric definitions introduced in section 2 .
the author is grateful to v. i. arnold for constant attention to this work , and to technische universitt graz for hospitality and excellent working conditions .
in this section we briefly introduce geometric aspects of ordinary continued fractions .
we start with general notions of continued fractions . for arbitrary sequence of real numbers @xmath1 the _ continued fraction _ is an expression @xmath2 denoted @xmath3 $ ] . in case of a finite sequence
we get some real number ( or sometimes @xmath4 ) . in case of infinite continued fraction
the expression means a limit of a sequence @xmath5)$ ] while @xmath6 tends to infinity .
notice that such limit does not exist for all sequences .
a continued fraction is called _ ordinary _ if @xmath7 is integer and the rest elements are positive integers .
a finite continued fraction is _ odd _ ( _ even _ ) if it contains odd ( even ) elements .
a rational number has a unique odd and a unique even continued fractions .
+ an irrational number has a unique infinite continued fraction .
we continue with several definitions of integer geometry .
a point is said to be _ integer _ if all its coefficients are integers .
the _ integer length _ of a segment @xmath8 is the number of integer points inside the segment plus one , denote it by @xmath9 .
integer sine _ of the angle @xmath10 is the index of the integer sublattice generated by the integer vectors of the segments @xmath11 and @xmath12 in the whole lattice , we denote in by @xmath13 . for more information on lattice ( in particularly integer )
trigonometry we refer to @xcite and to @xcite .
let @xmath14 be a cone with vertex at the origin .
take the convex hull of all integer points except the origin in @xmath14 .
the boundary of the described convex hull is a broken line together with one or two rays in case if there are some integer points on the edges of the cone .
the broken line in the boundary of the convex hull is called the _ sail _ of @xmath14 .
( in some literature the sail is the whole boundary of the convex hull , for us it is more convenient to exclude the rays from the definition of a sail . )
consider a positive real number @xmath15 .
denote by @xmath16 the cone with vertex at the origin and edges @xmath17 and @xmath18 .
the sail for @xmath16 is a finite broken line if @xmath15 is rational and one - side infinite broken line if @xmath15 is irrational .
so let the sail be a broken line @xmath19 ( @xmath20 ) .
denote @xmath21 for all admissible @xmath22 .
the sequence @xmath23 ( or @xmath24 ) is called the _ lattice length - sine sequence _ for the cone @xmath16 ( or _ lls - sequence for short _ ) . @xmath25 the connection of geometric and analytic properties of lls - sequence is introduced by the following theorem .
let @xmath0 and @xmath23 @xmath26or @xmath24@xmath27 be the lls - sequence for @xmath16 .
then @xmath28 \quad ( \hbox{or respectively $ \alpha=[a_0,a_1,\ldots]$}).\ ] ] this theorem is a reformulation of a proposition 1.7.a from @xcite for finite continued fractions and theorem 2.7.a from @xcite for the infinite continued fractions .
finite case is also a particular case of corollary [ geometry2 ] .
so we skip the proof here .
in this section we generalize geometry of ordinary continued fractions to the case of continued fractions with arbitrary elements .
we show a relation between odd or infinite continued fractions and broken lines in the plane having a selected point ( say , the origin ) .
we conclude this section with a few words about conditions for a broken line to be closed in terms of elements of the corresponding continued fraction .
further we use the following notation . for a couple of vectors @xmath29 and @xmath30 denote by @xmath31 the oriented volume of the parallelogram spanned by the vectors @xmath29 and @xmath30 . in this subsection
we give a natural geometric interpretation of an odd or infinite continued fraction with arbitrary elements .
it would be a broken line defined by the positions of the first vertex and the selected point @xmath32 , direction of the first edge , and the continued fraction .
so consider a continued fraction @xmath33 $ ] .
we are also given by the vertex @xmath34 , selected point @xmath32 , and the direction @xmath29 of the first edge .
we construct all the rest vertices @xmath35 inductively in @xmath22 .
_ base of induction_. for the second vertex we take @xmath36 where @xmath37 is defined from the equation @xmath38 . _ step of induction .
_ suppose now we have the points @xmath39 , for @xmath40 , let us get @xmath41 .
consider a point @xmath42 in other words @xmath43 is a point in the line @xmath44 such that the area @xmath45 equals 1 .
let @xmath46 finally the point @xmath41 is defined as follows ( see on figure [ induction ] ) @xmath47 @xmath48 let us now explain a geometric meaning of the elements of continued fractions in terms of characteristics of the corresponding broken line .
[ properties ] the following holds @xmath49 we prove this statement by induction in @xmath22 . _ base of induction .
_ from the definition of the point @xmath50 we get @xmath51 _ step of induction .
_ let the statement be true for @xmath52 , we prove it for @xmath22 . first , we verify the formula for @xmath53 : @xmath54 we have @xmath55 the step of induction is completed .
let us construct a broken line having the first vector @xmath56 , the direction @xmath57 , and the continued fraction @xmath58 $ ] .
then we have @xmath59 further we find the corresponding points @xmath43 and @xmath60 : @xmath61 finally we get @xmath62 now suppose we have a point @xmath32 and a broken line @xmath63 such that for any @xmath22 the points @xmath32 , @xmath35 , and @xmath41 are not in a line .
let us extend the definition of the lls - sequence for this data .
we use equalities of proposition [ properties ] to define the elements : @xmath64 we call the sequence @xmath65 the _ lls - sequence _ of the broken line with respect to the point @xmath32 , and @xmath33 $ ] the _ corresponding continued fraction_. [ conjugate ] let @xmath63 and @xmath66 be two broken lines with lls - sequences @xmath67 and @xmath68 respectively .
suppose the first broken line is taken to the second by some operator in @xmath69 with determinant equals @xmath37 .
then we have : @xmath70 the volume of any parallelogram is multiplied by @xmath37 , then the statement follows directly from formulas of proposition [ properties ] . for this subsection
we fix the point @xmath32 to be at the origin . consider a continued fraction @xmath71 $ ] as a rational function in variables @xmath72 .
this rational function is a ration of two polynomials with non - negative integer coefficients , denote them by @xmath73 and @xmath74 . actually the polynomials @xmath73 and @xmath75
are uniquely defined by the condition @xmath76\ ] ] and the condition that the coefficients of both polynomials are non - negative integer coefficients .
[ intcoef ] notice that the last condition is equivalent to the condition that the polynomial @xmath77 contains a monomial @xmath78 with unit coefficient .
[ geometry ] let @xmath63 be a broken line such that @xmath56 , and @xmath79 is collinear to the vector @xmath80 .
suppose its lls - sequence is @xmath81 .
then @xmath82 we prove this statement by induction in @xmath6 .
_ base of induction .
_ if the broken line is a segment @xmath83 with lls - sequence @xmath84 then @xmath79
. _ step of induction .
_ suppose the statement holds for all broken lines with @xmath22 vertices , let us prove it for an arbitrary broken line with @xmath85 vertex .
consider a broken line @xmath86 with lls - sequence @xmath87 .
let us apply a linear transformation with unit determinant taking @xmath50 to @xmath88 and the line @xmath89 to the line @xmath90 .
this transformation is uniquely defined by all these conditions , it is @xmath91 denote the resulting broken line by @xmath92 . by proposition [ conjugate ]
all the elements of the lls - sequence for @xmath92 are the same . by the assumption of induction we have
@xmath93 denote the coordinates of @xmath94 by @xmath95 and @xmath96 respectively .
then we have @xmath97 the polynomials satisfy @xmath98 notice that the polynomial @xmath99 has a unit coefficient in the monomial @xmath100 coming from @xmath101 . therefore ( see remark [ intcoef ] ) , @xmath99 coincides with @xmath102 and @xmath103 coincides with @xmath104 .
so we are done with the step of induction .
this concludes the proof of the theorem .
in particular we get the following corollary . in the classical case it forms the basis of geometry of ordinary continued fractions .
[ geometry2 ] let @xmath63 be a broken line such that @xmath56 , and @xmath105 .
suppose that the corresponding continued fraction is @xmath106 $ ] and @xmath107 .
then @xmath108 @xmath26if the corresponding continued fraction has an infinite value , then @xmath109.@xmath27 this corollary implies the following statement .
let @xmath63 and @xmath110 be two broken lines with @xmath111 , such that the vector of the first edges either have the same direction if @xmath112 or opposite otherwise .
suppose the corresponding continued fractions coincide : @xmath113=[b_0 , \ldots , b_{2m}].\ ] ] then the points @xmath114 , @xmath115 , and the origin @xmath32 are in a line .
consider an @xmath69-operator taking @xmath34 to @xmath88 and @xmath50 to the line @xmath90 . by proposition [ conjugate ] the continued fractions for both broken lines
are not changed .
hence by corollary [ geometry ] the points @xmath114 , @xmath115 , and the origin are in a line .
[ cc](_on closed curves . _ ) how to find that a certain continued fraction defines a closed curve ? from theorem [ geometry ] we see that a broken line defined by an lls - sequence @xmath81 with @xmath56 and @xmath83 being collinear to @xmath80 is closed if and only if @xmath116 so , _ these two polynomial conditions on the elements of the lls - sequence are necessary and sufficient conditions for the broken line to be closed_. notice that the condition @xmath117 can be rewritten in the following nice form @xmath118=0.\ ] ] the condition @xmath117 was introduced in @xcite for certain broken - lines with integer vertices .
let us study an example of broken lines consisting of three edges .
these curves are defined by continued fractions of type @xmath119 $ ] .
then the conditions for a broken line to form a triangle are as follows : @xmath120 ( see on figure [ examples ] . ) @xmath121$}=\frac{\displaystyle 0 } { \displaystyle 1 } & \hbox{$ [ 1,-2,2,-1/2,-4]$}=\frac{\displaystyle -1 } { \displaystyle 1 } \\ \end{array}\ ] ] there is one problem which is interesting in the frames of this section .
suppose we have a broken line and two distinct points @xmath122 and @xmath123 .
then we have two lls - sequences for the same curve with respect to @xmath122 and @xmath123 .
_ study the conditions on the initial data @xmath26i.e .
, lls - sequences , positions of the first points of the broken lines , and direction of the first vector@xmath27 that define congruent broken lines . _
now let us study what happens if we consider a curve as a broken line with infinitesimally small segments .
it turns out that the lls - sequence `` splits '' to a couple of functions which we call _ areal and angular densities_. we introduce the necessary notions and discuss basic properties of these functions .
in particular we show that the areal density is inverse to a velocity of a point defined by the second kepler law .
now we prove the theorem on finite reconstruction of a curve ( i.e. , in some small neighborhood ) knowing the areal density and a starting point .
this is analogous to the algorithm that finds a broken line by the elements of the corresponding continued fraction described in subsection [ algorithm ] .
the significant difference to the discrete case is that we do not need to know the angular distribution function .
notice that @xmath146 defines the oriented curvature .
therefore , if one knows the functions @xmath129 and @xmath130 then the curve is uniquely reconstructed until the time @xmath147 where the vectors @xmath148 and @xmath149 are collinear , or in other words where @xmath150 . consider a system of polar coordinates @xmath151 with the origin at the point @xmath32 . to get the curve
we should solve the system of differential equations : @xmath152 this system is equivalent to the union of the following two systems : @xmath153 let us say a few words about density functions and their broken line approximations .
let @xmath134 be a curve with arclength parameter @xmath155 $ ] and densities @xmath156 and @xmath157 .
for an integer @xmath6 consider a broken line @xmath158 such that @xmath159 .
let the corresponding lls - sequence be @xmath160 .
denote by @xmath114 and @xmath161 the following functions @xmath162 * ellipses and their centers . *
consider an ellipse @xmath167 with @xmath168 .
let @xmath32 be at the symmetry center of the ellipse i.e. at the origin .
then the areal and angular densities are as follows @xmath169 notice that here we get the constant function for the ratio : @xmath170 * ellipses and their foci . * as in the previous example we consider an ellipse @xmath167 with @xmath171 .
let now @xmath32 be at one of the foci , for instance at @xmath172 .
then the densities are as follows @xmath173 if we put the sun at the chosen focus and consider a planet whose orbit is the ellipse , then according to three kepler laws the planet will move with velocity @xmath174 at any @xmath126 . here
the constant @xmath37 is defined from the third kepler law : _ the square of the orbital period of a planet is directly proportional to the cube of the semi - major axis of its orbit , _ or in other words @xmath175 where @xmath176 is the period for our orbit , and @xmath177 and @xmath178 are respectively the period and the semi - major axis for the earth .
denote by @xmath179 the length of the ellipse , i.e. , @xmath180 since @xmath181 , we get @xmath182 then the densities for this spiral are as follows @xmath183 it is interesting to notice that for the spirals we have @xmath184 i.e. , the products are constant functions . | in this paper we introduce a link between geometry of ordinary continued fractions and trajectories of points that moves according to the second kepler law .
we expand geometric interpretation of ordinary continued fractions to the case of continued fractions with arbitrary elements . |
the search for supersymmetry is being carried out in many ways . in high energy collider experiments , such as the tevatron and lep ,
various measurements have yielded lower bounds on the sparticle masses , although no signal has yet been observed .
it is possible that we might need to wait for the next generation of colliders to further probe the spectrum of supersymmetric particles , although we may be able to get some interesting results through other types of experiments , such as the direct detection of dark matter particles .
it is well - known that in supersymmetric theories where r parity is conserved , the lightest supersymmetric particle ( lsp ) is stable , and it is a candidate for cold dark matter if neutral and colorless @xcite , as the lightest neutralino is @xmath13 . according to astronomical observations , most notably from the refurbished hubble space telescope , we know that most of the matter in the universe is invisible .
if we define @xmath14 , where @xmath15 is the critical density to close " the universe and @xmath16 the neutralino dark matter density , it is known that @xmath17 where @xmath18 and @xmath19 is the hubble parameter . since
recent hubble telescope observations indicate that @xmath20 @xcite , we see that @xmath21 is expected . with this large amount of dark matter in the universe
, it is expected that we should be able to detect the recoiling effect of the neutralino , as a dark matter particle , scattering off the target nuclei of certain types of detectors .
indeed , some experimental effort have been put into reality in this regard @xcite .
we consider two detectors @xmath22 and @xmath23 . the ge detector is currently being set up at stanford university , whereas the second one is still being discussed .
there are two points regarding the detector dependence of the scattering processes which are worth mentioning : ( i ) there are two processes through which the @xmath0 interacts with the partons in the protons and neutrons inside the target nucleus .
the first one is the spin - dependent ( s.d . )
incoherent scattering , which is non zero only when the target nucleus has non - zero spin .
the second one is the spin - independent coherent scattering which increases with the mass of the target nucleus ( @xmath24 ) .
these two processes do not interfere with each other , thus the total rate is just the direct sum of these two rates . for the two detectors considered here the spin - dependent contribution ( s.d . )
is always negligible compared with the spin - independent ( s.i . )
( ii ) because of these features , a detector with the highest possible atomic number is prefered : @xmath23 is the superconducting material with highest atomic number .
concerning the size of the detector , one would naively think that since one wants to have the largest event rate , one should build the largest possible detector .
however , this is not the case with the cryogenic experimental set up , whose operation is optimized for a 1 kg detector mass .
in what follows we investigate the parameter space of no - scale @xmath25 supergravity . for a recent review on the construction of these models see ref . @xcite .
the effect of several other direct and indirect experimental constraints have been discussed in refs .
these models have the features of universal soft - supersymmetry - breaking at the unification scale and radiative breaking of the electroweak symmetry , which is enforced by minimizing the one - loop effective potential . generally speaking , there are four free parameters that describe these models , once the top - quark mass is fixed .
they can be chosen to be : @xmath26 , where @xmath27 is the lightest chargino mass , @xmath28 is the ratio of higgs vacuum expectation values , and @xmath29 and @xmath30 are ratios of the usual universal soft - supersymmetry - breaking parameters . in string - inspired supersymmetry breaking scenarios two parameters are eliminated : in the _ moduli _ scenario @xmath31 , whereas in the _ dilaton _ scenario @xmath32 . in either case
there are only two free parameters left @xmath33 .
we also consider the `` strict '' version of these scenarios where the bilinear scalar coupling @xmath34 is also specified at the unification scale : in the ( strict ) moduli scenario @xmath35 , and in the ( strict ) dilaton @xmath36 . in this case
the value of @xmath28 can also be determined , and we are left with one - parameter models . compared with generic low - energy supersymmetric models like the minimal supersymmetric standard model ( mssm ) , these few - parameter models are better motivated and much more predictive .
the main inputs from these models to the dark matter detection calculation are : the lightest cp - even higgs mass @xmath37 , and the lightest neutralino mass @xmath38 .
the allowed ranges for these masses after all model constraints have been imposed are : @xmath39 and @xmath40 .
an important mass relation that one should keep in mind throughout our discussion is @xmath41 .
the feynman diagrams for the elastic scattering between the @xmath42 and the partons inside the nucleons in the target nucleus ( @xmath43 ) are shown in fig .
[ lspd_feynman ] . in order to calculate the scattering rate
we need to write down the effective lagrangian for this process .
if we assume that most of the dark matter halo surrounding our galaxy is made of elementary particle dark matter , then the typical dark matter velocity is @xmath44 , where @xmath45 is the speed of light . at these velocities
there is an energy transfer @xmath46 which is much smaller than the mass scale involved in the process ( @xmath47 ) .
therefore , we can use the effective lagrangian approach to describe the low - energy @xmath0-quark interaction , @xmath48 where @xmath49 and @xmath50 ( @xmath51 ) are the coupling coefficients for the @xmath52-exchange in the @xmath53-channel and the @xmath54-exchange in the @xmath55-channel ; only light quarks contribute in this case . also , @xmath56 are the coupling coefficients for the @xmath57-exchange @xmath53-channel and the @xmath54-exchange @xmath55-channel .
note that heavy quarks do not decouple in this case . among the various contributions to the scattering amplitude ,
the squark contribution is negligible since @xmath58 in these models .
the dominant contribution is from spin - independent coherent scattering due to ( cp - even ) higgs boson exchange , mostly the lightest one ( @xmath59 ) although the heavy higgs ( @xmath19 ) can have a significant sub - leading effect .
the total scattering rate is given by @xcite @xmath60 where @xmath61 is the target nucleus mass , @xmath16 is the neutralino dark matter relic density , @xmath62 is the average velocity of the neutralinos that hit the detector ( we take @xmath63 ) , and @xmath64^{2}\ .\ ] ] in the expression for @xmath65 , @xmath66 $ ] is a form factor with @xmath67 , @xmath68 the velocity dispersion , and the charge radius @xmath69 .
also , @xmath70 and @xmath71 are coefficients of proportionality in various hadronic matrix elements .
finally , the @xmath72 parameters for the dominant higgs exchange diagrams involve couplings of the higgs to neutralinos and to quarks : @xmath73 where @xmath74 is the higgs mixing angle , @xmath75 , @xmath76 , and the lsp is the admixture @xmath77 .
it can be seen that @xmath78 , thus the reason for proposing the @xmath23 detector .
also , explicit calculations show that for both @xmath79 and @xmath23 detectors , the s.i .
term is always much larger than the s.d .
term ( which we do not exhibit explicitly here ) .
.the fraction of parameter space in @xmath3 supergravity which can be explored via direct detection of dark matter in ge and pb detectors with a sensitivity of 0.1 or 0.01 events / kg / day . [ cols="^,^,^,^,^,^,^,^,^ " , ] ' '' ''
we have calculated the scattering rates for the neutralino dark matter particles in the two detectors mentioned above throughout the parameter space of the models .
the results are shown in the figs .
[ lspd_nsc ] , [ lspd_kl ] , [ lspd_strict ] for the different scenarios .
the following general features are noticed in all cases : * the rates drop quickly with growing chargino masses . as mentioned above ,
the lightest neutralino is a linear combination of gaugino and higgsino gauge eigenstates .
the intermediate states for the dominant spin - independent scattering process are the higgs bosons , which couple to a product of the gaugino and higgsino components of the neutralino ( see eq . ( [ cq ] ) ) . on the other hand , when the chargino mass grows the neutralino approaches the limit of a pure bino , thus greatly reducing its coupling to the higgs bosons , and the scattering rate drops sharply .
* there are two noticeable dips in the rates for @xmath80 . for @xmath81 , in the calculation of the relic density of the neutralinos ( which we perform following the methods of ref .
@xcite ) , the @xmath52 or @xmath59-pole is encountered in the neutralino pair - annihilation , and the resulting neutralino relic density is highly suppressed .
this implies that little neutralino dark matter is expected in the halo . to account for this possibility , in eq .
( [ r ] ) we take @xmath82\ , 0.3 { \gev}{\rm cm}^{-3}\ , \ ] ] since one expects that @xmath83 @xcite . in the figures ,
the ordering of the poles with increasing @xmath27 is @xmath84 for @xmath85 , and @xmath86 for @xmath87 ( and in fig .
[ lspd_strict ] ) . *
the rates have a non - trivial @xmath28 dependence ( only @xmath88 are shown in the figures ) .
the @xmath72 coefficients in eq .
( [ cq ] ) depend on @xmath28 , and so does @xmath89 , which increases with @xmath28 . for the ge detector , which is the one
being set up right now and should start producing data in the near future , with the realistic experimental sensitivity of 0.1 event / kg / day , only a small portion ( @xmath90 ) of the parameter space would be probed in all scenarios ( see table [ table ] ) .
these fractions have been determined by computing the rates for _ all _ points in parameter space ; only selected values of @xmath28 are shown in the figures .
there are two ways to improve the reach into parameter space .
first , one could improve the sensitivity by reducing the background . if the sensitivity of the ge detector could be enhanced to 0.01 event / kg /
day , a larger fraction of the parameter space could be probed ( @xmath91 ) , as seen in the figures and quantified in table [ table ] .
there has also been some discussions about the possibility of building a @xmath23 detector , which is an improvement over the ge detector because the scattering rate scales about linearly ( @xmath92 ) with the atomic number of the target nucleus , , an improved sensitivity by increasing the signal in the detector . with the 0.1 event / kg / day sensitivity the pb detector should probe a larger ( @xmath93 ) fraction of the parameter space , compared to the ge detector with the same sensitivity ( see table [ table ] ) .
of course the best experimental scenario would be a @xmath23 detector with the better sensitivity ( 0.01 events / kg / day ) .
in such best case scenario , for @xmath85 and @xmath94 , chargino masses as high as @xmath95 could be probed in the moduli ( dilaton ) scenario .
the reach is somewhat lower for larger values of @xmath28 .
note also that in strict no - scale @xmath3 dilaton scenario , an experimental upper bound on @xmath96 would immediately provide a lower bound on @xmath27 .
it has been recently pointed out @xcite that experimental limits on @xmath97 may disfavor some regions of parameter space where @xmath96 is particularly enhanced .
the computations of @xmath97 in supergravity models in ref .
@xcite , in particular for no - scale @xmath3 supergravity , indicate that @xmath98 is an important constraint only for @xmath99 , and even in that case only for @xmath100 .
therefore , even assuming that the @xmath98 constraints are rigorous , there should still be a large portion of the parameter space accessible to direct searches of dark matter .
we conclude with the following comment . at lepii
, chargino masses as high as 100 gev should be readily detectable in all the scenarios which we have considered here @xcite , and yet , direct detection of the lsp with @xmath101 may take a while to occur .
indirect detection of dark matter via upwardly moving muon fluxes in neutrino telescopes may taken even longer to happen , as explicit calculations show in the models we considere here @xcite and more generally in model - independent analyses @xcite .
this work has been supported in part by doe grant de - fg05 - 91-er-40633 .
the work of x. w. has been supported by the world laboratory .
we would like to thank d. caldwell for information on the status of the experiments .
99 j. ellis , j. s. hagelin , d. v. nanopoulos , k. a. olive and m. srednicki , .
w. freedman , , nature * 371 * ( 1994 ) 757 .
see , p.f .
smith and j.d .
lewin , phys .
* 187 * ( 1990 ) 2093 ; b. young , , in proceedings of the xxvi international conference on high energy physics ( dallas , 1992 ) , ed . by j. r. sanford ( aip conference proceedings no .
272 ) , p. 1260 . , , and a. zichichi , prog . part . nucl . phys .
* 33 * ( 1994 ) 303 . , ,
g. park , x. wang , and , .
, , and a. zichichi , ( hep - ph/9408345 ) .
w. goodman and e. witten , ; k. griest , ; ; r. barbieri , m. frigeni and g.f .
giudice , ; m. srednicki and r. watkins , ; g.f .
giudice and e. roulet , ; g. gelmini , p. gondolo , and e. roulet , ; m. kamionkowski , ; r. flores , k. olive , and d. thomas , ; j. ellis and r. flores , ; ; ; m. drees and m. nojiri , ; a. bottino , , astropart . phys . * 2 * ( 1994 ) 77 ; v. bednyakov , h. klapdor - kleingrothaus , and s. kovalenko , and hep - ph/9401262 ; r. arnowitt and p. nath , cern - th.7362/94 ( hep - ph/9408226 ) and cern - th.7363/94 ( hep - ph/9409301 ) . , , k. yuan , ; s. kelley , , , h. pois , and k. yuan , ; jl , , and k. yuan , .
see , e. kolb and m. turner , _ the early universe _
( addison - wesley , 1990 ) .
f. borzumati , m. drees , and m. nojiri , kek-400 ( hep - ph/9406390 ) ; r. arnowitt and p. nath , cern - th.7363/94 ( hep - ph/9409301 ) .
, , x. wang , and , ( hep - ph/9406427 ) ( to appear in phys
r. gandhi , , , k. yuan , and a. zichichi , .
m. kamionkowski , k. griest , g. jungman , and b. sadoulet , hep - ph/9412213 . | we compute the scattering rates for the lightest neutralino @xmath0 in the forthcoming germanium @xmath1 detector and a proposed lead detector @xmath2 , within the framework of @xmath3 supergravity .
we find that in only a small portion ( @xmath4 ) of the parameter spaces of this class of models , are the rates in the germanium detector above the expected initial experimental sensitivity of 0.1 events / kg / day .
however , a much larger portion ( @xmath5 ) of the parameter spaces could be probed with an improved background rejection capability ( 0.01 events / kg / day ) and/or a more sensitive detector @xmath6 .
s-6pt/ h-7.5pt/ # 1re#1 # 1im#1 # 1tr#1 = 11 # 1#2 # 1#1| # 1| # 1 # 1#1 # 1#1 versim#1#2 # 1@xmath7 # 1@xmath8 11 # 1#2 /#1#1 - 6pt/ # 1#2#3nucl .
phys .
b * # 1 * ( 19#2 ) # 3 # 1#2#3phys . lett .
b * # 1 * ( 19#2 ) # 3 # 1#2#3b * # 1 * ( 19#2 ) # 3 # 1#2#3phys .
rev .
d * # 1 * ( 19#2 ) # 3 # 1#2#3phys .
rev .
lett . *
# 1 * ( 19#2 ) # 3 # 1#2#3phys . rep . *
# 1 * ( 19#2 ) # 3 # 1#2#3mod
. phys .
lett . a * # 1 * ( 19#2 )
# 3 # 1#2#3int .
j. mod .
phys .
a * # 1 * ( 19#2 ) # 3 # 1texas a & m university preprint ctp - tamu-#1 # 1#2#3ann .
rev .
. astrophys . *
# 1 * ( 19#2 ) # 3 # 1#2#3ann .
rev .
nucl . part .
sci .
* # 1 * ( 19#2 ) # 3 6.0 in 8.5 in -0.25truein 0.30truein 0.30truein + ctp - tamu-63/94 + act-21/94 + hep - ph/9412237 + 0.5 cm * direct detection of dark matter in su(5)@xmath9u(1 ) supergravity + * 0.4 cm _
@xmath10center for theoretical physics , department of physics , texas a&m university + _ _ college station , tx 778434242 , usa + _ _ @xmath11astroparticle physics group , houston advanced research center ( harc ) + _ _ the mitchell campus , the woodlands , tx 77381 , usa + _ _ @xmath12cern theory division , 1211 geneva 23 , switzerland + _ 0.5 cm + ctp - tamu-63/94 + act-21/94 + november 1994 |
in this paper we consider transformation : @xmath7 generated by the action of phase flow of the vector field @xmath8 : @xmath9 in the case when coordinate functions of vector fields are the polynomials of finite degree .
calculation of the action of phase flows is equivalent to solving of the system of odes . in the study of odes
, we often use pertrubation series methods @xcite . in such cases
we decompose right hand side of an equation into the series , take linear part , and consider all the other as small in some neighborhood : @xmath10 here @xmath11 is a linear operator , @xmath12 - some small parameter , @xmath13 - some function . in cases ,
when it is not confusing , we will omit indexes , which mark the components of the vectors and tensors .
a solution of the system is presented as series by powers of parameter @xmath12 : @xmath14 if it converges in some neighborhood of initial point .
a term behind of @xmath15 could be calculated from a linear differential equation : @xmath16 whose solution is delivered by matrix exponent : @xmath17 @xmath18 with linear algebra , we could calculate exponent in the basis , where basis vectors are the eigenvectors of @xmath11 .
but in case , when we want to keep nonlinear terms , such algebraic background is not developed yet @xcite .
+ an objective of this paper is to investigate one possible way to create such algebraic framework , based on the diagrams , consisting of contracted @xmath19-tensors .
it is instructive to consider this technique on a discrete dynamic example .
for simplicity sake , let s consider a map : @xmath20 where @xmath21 - is a homogeneous polynomial of degree @xmath22 . in indexes ,
we can rewrite it as : @xmath23 where a summation by the repeating indexes is introduced .
the next iteration of the map can be gotten by the next iteration substituting instead of @xmath24 : @xmath25 for convenience contraction notation , we introduce the following diagrams : + where the below notations are introduced : + a number of operators , which diagram consists of we call * an order * of diagram , and denote @xmath26 .
we will omit indexes , arrows over the edges of the diagrams , and vertexes signatures in cases , when it is not confusing .
summation is to be done by the repeating indexes .
the arrows direction is always from right to left .
+ for recurrence with arbitrary non - linearity on the r.h.s : @xmath27 evolution operator looks like : @xmath28 where summation goes over all possible diagrams of order @xmath29 , which we can get by contraction of operators @xmath30 .
a calculation of such sum in general case is not a subject of this papers discussion .
in this section and until the very end we consider general systems of differential equations with the polynomials in the right side .
we will find ansatz for a generic solution and make some estimations for its converging radius .
let s consider a system of autonomous differential equations of general form , with the polynomials in r.h.s : @xmath31 for simplicity , we consider homogeneous polynomials of the same degree in r.h.s . generalization for sum of few homogeneous components is trivial , as it will become clear later .
@xmath32 @xmath33 let s transform this equation into a recurrence . for that
, let s split a time interval @xmath6 by @xmath34 similar intervals @xmath35 , and rewrite equation as : @xmath36 to get an explicit solution , we should fix @xmath6 and calculate limit , if it exists : @xmath37 here a composition means , that with every upper index of previous composite - multiplier , we contract lover index of the next one . , scaledwidth=65.0% ] we could get the next iteration by contraction of operator @xmath38 with every free edge on the right side of every diagram , and expanding by linearity .
the @xmath39 iteration will look like : @xmath40 where @xmath26 - order of diagram @xmath41 , @xmath42 - coefficient near diagram @xmath41 for iteration @xmath43 , @xmath44 means summation over all diagrams of order @xmath41 .
but we need to get limit @xmath45 , so evolution operator looks like : @xmath46 where @xmath47 - asymptotic coefficient .
we could get its value by direct calculation , but we will use some trick here . for r.h.s . with the several homogeneous components
an answer will have the same form , but in the diagrams there will also be forks , corresponding to all homogeneous components of r.h.s .
( in the further text , operators in diagrams would be sometimes referred as the vertices ) .
the trick is that we substitute a solution : @xmath48 in the equation ( @xmath49 - operator of order @xmath22 ) : @xmath33 @xmath50 now , if we compare the coefficients before the same diagrams , we get : @xmath51 where the diagram @xmath41 arises from the operator @xmath49 , to the free edges of which we have contracted @xmath52,@xmath53 ... @xmath54 : in recurrence for @xmath55,scaledwidth=50.0% ] in this formula , we could substitute @xmath56 in the same way , and so on .
we repeat this process , until we come to the trivial diagram , which consists of one identical operator , a coefficient before which is 1 .
thus , the coefficients look like : @xmath57 where product is calculated by the sub - diagrams , which we could get by taking out all sub - trees , which originate from the each vertex , and have the same right , end , edges as @xmath41 .
generally : @xmath58 where @xmath59 - some diagram , which is also constructed from the basic vertexes .
we will use the same notation in the further text .
in such way , a general form of evolution operator is : @xmath60 which really looks like a matrix exponent in which we sum up by all branching degrees of the diagrams .
it is the key formula of this text .
the folloiwng inequality holds for euclidean norm of contraction of two operators : @xmath61 thus , for evolution operator for equation with homogeneous non - linearity of degree @xmath22 we could make an estimation : @xmath62 @xmath63 where @xmath64 .
so , we can also estimate a convergence radius of the series : @xmath65{c^p_\sigma \parallel d \parallel^p } } = \dfrac{1}{\parallel d \parallel } \lim\limits_{p\to\infty}\dfrac{1}{\sqrt[p]{c^p_\sigma}}\ ] ] we get coefficients @xmath66 from the one dimensional equation of such kind : @xmath67 @xmath68 after integration , we get : @xmath69 @xmath70{1-\dfrac{\alpha x_0^{s-1}}{s-1}t}}\ ] ] @xmath71 space is one dimensional , therefore : @xmath72 so , all diagrams with the same number of the vertices will have same values . from solution ( [ solution_1 ] ) we could get the coefficients : @xmath73 after substitution of ( [ c_sigma_k ] ) into ( [ convergence_estimation_1 ] ) , we get @xmath74 in the same way we could get estimation for convergence radius for more complicated r.h.s . , which consist of more quantity of the homogeneous components .
difficulty here is , that during the process of calculating of @xmath75 we need to solve algebraic equations of higher degree , to calculate @xmath76 in the one dimensional case .
in this section we consider how , if we know solution for a linear system , to built non - linear perturbations theory over it .
we consider ansatz substituting procedure as explained in the previous section .
estimations for convergence radius for a non - resonance cases are also presented . as to classical results on dynamic in the neighbourhood of the fixed points see , for example,@xcite or @xcite . if we decompose r.h.s .
of equation near the not degenerated fixed point , equation gets the form : @xmath77 where @xmath11 - some matrix , @xmath21 - finite degree polynomial .
let s do substitution , which is called in physics an interaction representation : @xmath78 @xmath79 @xmath80 @xmath81 now , let s rewrite it as an integral equation , and iterate : @xmath82 @xmath83 and so on . here , all calculations are on the level of formal series
, a convergence question will be discussed later .
+ so , a solution looks like a sum by all diagrams , where the vertexes connected with exponents - propagators , are orderly integrated by all intermediate times . in case ,
when a linear operator is diagonalizable , the propagators looks like : @xmath84 where @xmath85 - eigenvalues of matrix @xmath11 .
specific of the these series , in compare with the previous one , is that here the indexes inside of the diagram play an important role - eigenvalues in intermediate exponents depend on them .
so , if we want to rewrite answer as a sum over all diagrams , we should sum up by all skeleton diagrams - every vertex corresponds to one component of nonlinear operator @xmath86 only ( for example , @xmath87 and @xmath88 calculated separately ) .
thus , ansatz looks like : @xmath89 where @xmath90 - some coefficients , depending on @xmath91 , @xmath92 - vector , which consists of eigenvalues , @xmath93 . in case ,
when operator is not diagonisable , or exist resonances ( @xmath94 ) , terms like @xmath95 would also arise .
let s consider , for simplicity , an equation in @xmath96-dimentional space : @xmath97 generalization is trivial again .
let s look for its solution in the following form : @xmath98 where summation goes by all diagrams , which consist of all monomial operators , e.g. different terms in summation in contraction goes with the different coefficients @xmath99 , @xmath41 - skeleton diagram .
we suppose , that frequencies is not resonance : @xmath100 . here
it is important , that on right side of the diagrams not @xmath101 is contracted , but some coefficients @xmath102 , which are not known , but depended on initial conditions . substituting ( [ ansatz_interaction_rep ] ) in ( [ equation_non_resonance_case ] ) , we get : @xmath103 @xmath104 where @xmath105 - unit vector along @xmath106-th axis .
the coefficients before the similar exponents and diagrams are : @xmath107 @xmath108 @xmath109
now , we can substitute every @xmath110 for its decomposition , and repeat this process until the diagrams without vertices remain . a coefficient before a term , without vertices is : @xmath111 @xmath112 thus , only non - zero coefficient can be put before the term , where @xmath113 .
now , coming back along the diagram from its right side to the left one , we see that in every vertex only one exponent is left , which is defined by the indexes on its right edges .
@xmath114 @xmath115 where product is done by all diagrams vertexes , and a nested sum - by all free edges , which are on the right side from vertex @xmath116 .
symbol @xmath117 means index , which corresponds to edge , @xmath118 - index , which corresponds to the left edge , which enters to diagram . for the further convenience
, we could rewrite it as : @xmath119 where product is calculated by the sub - diagrams , which we could get by taking out all sub - trees , which originate from the each vertexes ( as in ( [ coefficient_general ] ) ) , and @xmath120 means a sum of all eigenvalues , which correspond to the edges , which originate from the vertex ( or diagram ) , with the correct sign : @xmath121 this function is good according to the contraction ( denoted with @xmath122 ) : @xmath123 the coefficients @xmath124 are defined from initial conditions through equation : @xmath125 so , dynamic in the neighborhood of a fixed point looks like : @xmath126 where @xmath127 is the function , whose coefficients we just calculated , and the variables @xmath102 are multiplied by the corresponding exponents .
the constants dependency on initial conditions can be expressed as : @xmath128 where @xmath129 - a function inverse to @xmath127 .
its explicit form can be obtained , if we suppose , that it could be represented as the series by diagrams , with some coefficients : @xmath130 @xmath131 @xmath132 if we do same procedure with @xmath133 , as in the previous cases , we get : @xmath134 where @xmath135 - all possible ways to split diagram on sub - diagrams over edges , @xmath136 - number of the sub - diagrams in partition , @xmath137 - product by all sub - diagrams in partition . here
we clearly see a small denominators problem - even in non - resonance case , denominators @xmath138 can be arbitrary small . from diophantine approximations theory @xcite , we know , that for the majority of @xmath139 ( except lebesgue measure zero ) the following estimation is applicable : @xmath140 for simplicity , let s consider a case , when diagrams consist only of operators with @xmath22-forks .
asymptotically : @xmath141 @xmath142 @xmath143 @xmath144 @xmath145 estimation for convergence radius : @xmath146{\sum\limits_{|d|=n}(d!)^{\nu}}\right)^{-1}\ ] ] in the same way , we could get one for @xmath129 : @xmath147{\sum\limits_{|d|=n } \sum_{\pi\in\mathcal{p}(d)}(-1)^{\#(\pi)+1}\prod\limits_{q\in\pi}(q!)^{\nu } } \right)^{-1}\ ] ]
in this section , we will consider perturbations theory in the case of presence of resonances . to calculate series for solution
, we will develop new approach , based on the direct perturbation series coefficients calculation .
here we come up to the calculation of the evolution operator directly from the limit transition .
we represent @xmath148 decomposed along the time - when we calculate composition , we assign composition multiplier to the time , when it acts , like an infinitesimal transformation , and consider the terms in particular order : let s consider an equation : @xmath77 where @xmath149 - again , for simplicity , homogeneous operator - generalization for larger number of homogeneous components is obvious .
evolution operator looks like : @xmath150 let s suppose , we have expanded all brackets .
let s consider the terms in order of increasing of number of occurrence of @xmath21 in them .
the first term does nt have @xmath21 in it at all . calculating composition from left to the right
, we get : @xmath151 now , let s consider the terms , with one occurrence of @xmath21 in them .
let s consider all terms , were @xmath21 appear on @xmath43-th position .
thus , summation by all terms with different number of appearing of @xmath11 gives us the ordinal exponents before , and after @xmath43-th position : @xmath152 where the second composition is got on all edges . in a limit transition @xmath153 ,
a sum by @xmath43 should be replaced by the below integral : @xmath154 @xmath155 @xmath156 and so on . here
we should be careful with a times order @xmath157 . here
, we consider separately all terms of non - linearity , assuming that linear part is diagonalised , without jordan cells , e.g. : @xmath158 since we are nt interested in the particular coefficients values , but in existence of the terms with some particular indexes only , we represent ( [ example_resonances ] ) in the form : @xmath159 contraction of indices gives : @xmath160 using such approach , we could look at the diagrams in evolution operator from a little bit different point of view - it has a sum of all possible terms , generated by such substitutions inside of it .
existence of substitution in the set of generators implies an existence of corresponding component in the r.h.s . of equation .
with every such a diagram we have an ordered integration by time .
thus , we will have the terms like : @xmath161 if a factor multiplied by @xmath6 in the exponent is zero , a term with @xmath162 is present in solution . but this relates not to such diagram only , but to all diagrams , which have such sub - diagram . in some specific cases , it could be reduced , but this is not the case for the general coefficients . +
a criterion of existence of resonance terms is existence of solution in natural numbers ( or with one @xmath163 - corresponding to the left edge ) of equation : @xmath164 even in case of three @xmath165-valued eigenvalues it is a non - trivial question .
but let us imagine , that we have solved this equation .
the next necessary question is if such term would be delivered by generators .
necessary condition is existence of solution of matrix equation in @xmath166 : @xmath167 where @xmath11 - matrix , which columns are the generators , with multiplicity of every index ( e.g. @xmath168 or @xmath169 ) in every cell , @xmath170 - solution , for resonance set of eigenvalues . but even if this equation has a solution , still there is an open question , if we are really able to build such diagram from the generators .
+ in terms of function @xmath91 defined in the ( [ definition_lambda ] ) , every resonance is corresponding to the solution of the equation : @xmath171 for some d , generated by r.h.s of equation .
+ * example .
* let s consider two - dimensional system , with commensurable eigenvalues @xmath172 , @xmath173 , @xmath174 .
let s assume , for example , that in a diagonal basis , its non - linearity has the terms @xmath175 , @xmath176 .
thus , a criterion of existence of resonance terms is : @xmath177 for some positive @xmath178 .
such diophantine systems are well investigated in the numbers theory @xcite .
+ * example . *
numerical simulation + verification was done for systems in @xmath179 with the pure imagine eigenvalues @xmath180 .
the results are presented in an appendix .
+ the existence or absence of resonance terms could change perturbations theory completely .
ansatz for system with resonance terms would look like : @xmath181 getting of coefficients from the direct ansatz substitution is technically very hard , so we come to resonances problem from the ordered integration point of view . for the beginning ,
let s calculate an integration on one vertex - something like ( [ example_ordered_exp ] ) .
we could rewrite it in the diagonal basis and regroup eigenvalues behind coinciding times : @xmath182 @xmath183 if this integration is the step in the calculations in the bigger diagram , then we subdivide further calculations onto two scenarios : 1 .
we take the first term in the sum .
eigenvalues are accumulating - it means , that when we will calculate next diagram , which action is located in time @xmath184 , then eigenvalue @xmath85 will be canceled , and we could think about edge @xmath106 as it was replaced with edges @xmath185 and @xmath29 , with the corresponding exponents .
we take the second term in the sum .
eigenvalues accumulation is interrupted , which means , that when we will calculate next diagram , which action is located in time @xmath184 , then eigenvalue @xmath85 will not be canceled , and all previous part of diagram is died out - indexes @xmath185 and @xmath29 will not take part in the further calculations , and we need to multiply the diagram on @xmath163
. we will interpret further edge @xmath106 as the beginning edge with @xmath85 in exponent only . and
the same algorithm is applied for the all other vertexes .
now , we are ready to write down an answer for a non - resonance case : 1 .
the first summation goes over all diagrams @xmath186 , which could be constructed from the r.h.s . of equation .
the second summation goes over all partitions @xmath187 of diagrams .
each edge , where diagram is splitted into two sub - diagrams , corresponds to taking of the second term in the ( [ ordered_exponent_one_vertex ] ) , and each solid sub - diagram corresponds to the first terms , to the integration without interruptions. 3 . in each term in the nested summations we have @xmath188 .
each sub - diagram @xmath189 is multiplied on the @xmath190 and on the coefficient : @xmath191 , which comes from the integration on the each vertex of the solid sub - diagram .
the whole diagram is multiplied by the additional @xmath190 .
general degree of @xmath190 behind each diagram is the same , as the number of the sub - diagrams in the partition , plus @xmath192 .
let s @xmath193 be the only sub - diagram , which left edge is not contracted .
the diagram @xmath41 is multiplied on the @xmath194 , corresponding to the right edges of the sub - diagram , which had not died out until the very end of the integration along diagram @xmath41 . in general , that all gives us a result coinciding with the one , obtained from the direct ansatz substitution .
we could introduce a normalized diagram , to simplify the answer : @xmath195 @xmath196 @xmath197 where @xmath198 has been defined in ( [ coefficient_resonance ] ) , and @xmath199 means , that we took diagram s partition , normilize each sub - diagram , and contract normilized sub - diagrams in the same order , as in the original diagram .
+ in the case of presence of resonances , we have got much more interesting picture . in the vertex ,
were resonance was born , we have simply : @xmath200 ( as it was mentioned in the previous subsection ) .
we could consider this moment like a specific interrupting in the meaning , that exponent accumulation is interrupted , or as a birth of the resonance .
in every diagram there could be several such moments .
@xmath184 , which has been born in the resonance will be preserved in the further integration , if we do not choose dying out alternative .
such resoncnce lines ( integration with @xmath6 , which getting start on the some resonance diagram s initial vertex ) could meet each other , and degree of @xmath6 would increase .
integration on a vertex with the arbitrary degree of @xmath6 will look like : @xmath201 @xmath202 where @xmath203 - ordinal thing , which we have seen in the previous calculations .
so , we could also use normalization , but with the small changes - if the sub - diagram @xmath204 is resonancial , we should nt divide it on the last coefficient @xmath205 - because it is zero . to interpret terms in the brackets we should consider each resonance line which comes to this vertex separately , like a new degree of freedom , and look to each @xmath6-degree decreasing like a die of one of these resonance line. as in the non - resonance case , in case of dying out of each resonance line we should multiply expression by the @xmath206 .
factor @xmath207 behind @xmath208 means the numbers of ways , how we could chose @xmath43 resonance lines from @xmath22 for the dying out. so , in some meaning , overlaped resonance lines behaves independently - every resonance line in the integration behave , like an independent one , and their evolution scenarios are multiplied. each resonance line in diagram @xmath41 , which takes part in the integration between two resonance vertexes @xmath209 and @xmath210 , and die between them , will give a multiplier : @xmath211}\lambda_v^{-1}\ ] ] where @xmath212 $ ] - set of all vertexes , lying between @xmath209 and @xmath210 , @xmath213 - eigenvalues sum , corresponding to the vertex @xmath116 .
now a case of nested resonance diagrams is left unclear .
let @xmath41 be some diagram , @xmath214 - some resonance sub - diagram : @xmath215 .
we call @xmath216 nested resonance sub - diagram , if @xmath217 and @xmath218
. there could be several nested resonances in the @xmath204 on the same or different level of embedding .
we will call diagram ( not obviously resonance ) * irreducible * , if it havent got nested sub - diagrams . + every reducible resonance diagram is a contraction of irreducible .
indeed , let @xmath219 be reducible resonance diagram , @xmath218 . thus : @xmath220 if @xmath21 - reducible , proposition is proved .
if not , we could repeat this process .
such an algorithm of decomposition is finite , because @xmath221 .
+ in the vertex , where overlaped resonances come to the new vertex , giving birth to a resonance , we will have an integral : @xmath222 here we have coefficient @xmath223 , which breaks our superposition principe for resonance lines. now , we are ready to write down a general algorithm for the resonance case
. we subdivide it onto three parts .
the main algorithm is : 1 .
first summation goes over all irreducible non - resonance ( normalized ) diagrams @xmath224 , which could be constructed from the r.h.s . of equation .
2 . in each term of the summations we have @xmath225 .
second summation goes over all possible ways of contraction of free bottom indexes of @xmath224 with @xmath226 - deformed initial conditions ( see the third algorithm ) or @xmath227 - the resonance sub - diagrams coefficients ( see the second algorithm ) , for all alowed @xmath22 .
4 . multiply each term on the @xmath228 according to every @xmath226 contracted with bottom indexes of @xmath224 .
5 . multiply each term by the @xmath229 , where @xmath116 - every vertex , where @xmath227 is contracted with the bottom indexes of @xmath224 ; @xmath106 - initial vertex of @xmath224 : such a vertex , that contains only upper index of @xmath224 .
on each free edge @xmath116 of @xmath224 we could collect all expressions contracted with it , and get : @xmath230 where @xmath231 - summation over all resonance diagrams , @xmath232 - number of irreducible resonance sub - diagrams of @xmath233 .
but , unfortunately , @xmath234 depends on the diagram @xmath41 and even on its free edge @xmath116 , which it has been contracted with .
+ in the second algorithm , we calculate pure resonance contributions @xmath235 . for each resonance diagram @xmath236 : 1 .
write down @xmath237 , where @xmath238 - the deformed initial conditions .
2 . decompose @xmath233 into contraction of irreducible resonance diagrams @xmath239 .
summation goes over all possible scenarios of @xmath240 resonance lines dying out , where @xmath241 - number of irreducible resonance diagrams in @xmath233 .
4 . each scenario describing by choosing of @xmath242 s : each resonance line getting its start in the initial vertex @xmath243 of some @xmath244 s and goes to the initial vertex of diagram @xmath224 . for each resonance line ,
we need to choose another vertex @xmath242 ( which is also an initial vertex for some @xmath245 ) on the way from @xmath243 to @xmath106 , for dying out. 5 . multiply diagram by the @xmath246 6 .
multiply each initial vertex of @xmath244 s by the @xmath247 , where @xmath248 - the number of resonance lines , which passing this vertex . and in the third algorithm
, we calculate deformed initial conditions
@xmath249 : 1 .
first summation goes over all diagrams @xmath250 , which could be constructed from the r.h.s . of equation .
second summation goes over all partitions @xmath251 .
3 . for each @xmath252 , write down normalized @xmath224 .
4 . decompose @xmath224 into contraction @xmath253 , where @xmath254 - irreducible non - resonance diagram , @xmath255 - resonance diagrams .
summate over all possible ways of choosing of @xmath256 s ( @xmath256 have the same role as @xmath22 in the @xmath235 ) .
multiply each @xmath257 by the coefficient , similar to the one , obtained in the previous algorithm , for some @xmath256 .
multiply each term by the @xmath258 , where @xmath259 - every vertex , where some @xmath260 is contracted with a bottom index of @xmath254 ; @xmath106 - initial vertex of @xmath254 , @xmath256 - number , chosen for a diagram @xmath260 , defined below .
we can interpret each irreducible resonance diagram as a new effective degree of freedom of the system , dynamic along which is simple trivial @xmath6 , and solution of the system in some neighborhood of the fixed point would look like : @xmath261 where each @xmath6 corresponds to some irreducible resonance diagram , @xmath262 - deformed initial conditions . or , if we decompose the map : @xmath263 where @xmath264 - number of irreducible resonance diagrams ( not obviously finite number ) , dynamic in @xmath265 would look like : @xmath266 and dynamic in @xmath267 would look like : @xmath268 resonance part of linearized phase space definitely have finer structure .
vector fields space , which we have investigated , constitutes a lie algebra relating to a commutation operation : @xmath269 @xmath270 = \sum ( u^i\partial_i v^j - v^i\partial_i u^j)\partial_j\ ] ] evolution operator is an action of some subgroup of a group @xmath271 , gotten by exponentiation , on @xmath272 .
a problem of the formalism developed here for initial subgroup investigation , is that evolution operators action is defined locally , by series .
it s not clear how to get global action properties on the whole @xmath272 .
the global properties of the lie group reconstruction by exponential elements is a complicated task .
+ the interesting properties of the considered algebras are that their universal enveloping algebra has branching algebra structure - composition of pairs of elements defined not uniquely , e.g. : @xmath273 or @xmath274 we could contract a pair of elements by different indexes , and get different result
. the lie algebra structure could be restored , by introducing a commutator : @xmath275=\sum\limits_{k=0}^{a}u^i_{i_1 ... i_a}v^{i_k}_{i_{a+1} ... i_{a+b } } - \sum\limits_{k=0}^{b } v^i_{i_1 ... i_b } u^{i_k}_{i_{b+1} ... i_{a+b}}\ ] ] which we get from the vector fields commutator ( or from quadratic part of evolution operators commutator ) .
if we continue an analogy with linear algebra , we could introduce something like a linear operator product : @xmath276 and the operators commutator gets a traditional form : @xmath275=uv - vu\ ] ] unfortunately , this supporting product is not even associative ( evolution operators composition is , obviously , associative ) .
we also could make an assumption , that according to this product , an evolution operator could be simply represented as : @xmath277 a direct proof of this fact is related to some combinatorial difficulties .
+ in the same way , as we could consider the matrix lie algebras in abstract way , forgetting about their matrix nature , coming back to it by representation - branching algebra is associative algebra , a product of pair of the elements in which could be done in the different ways .
representation of branching algebra should be done not to @xmath278 , but to @xmath279 : @xmath280 where @xmath281 .
if we act with @xmath11 on @xmath8 , all interesting non - linearity is coming from a lifting to the diagonal operator : @xmath282 @xmath283 so the action of the algebra @xmath11 on @xmath8 is delivering by an composition : @xmath284 some freedom in @xmath285 choosing also exists : it is easy to see , that exists @xmath286 such that @xmath287 .
+ actually , the considered structure has not linear , but rather affine nature .
to illustrate this idea , let s consider an example .
+ * example .
* we consider sub - algebra @xmath288 of polynomial vector fields algebra , which has non - degenerated zeroes in the points @xmath289 : @xmath290 @xmath291 when we decompose it in the neighborhood of non - degenerated zeroes , it has the following form : @xmath292 @xmath293 where @xmath11 and @xmath294 - some matrices , @xmath295 and @xmath296 - polynomials .
the commutators of these fields would look like : @xmath270=z(ab - ba)\partial + f(z)\partial\ ] ] where @xmath297 - some new polynomial .
a linear part of the vector field in the neighborhood of every point commutes in the same way , independently of the chosen coordinates .
thus , if we consider an algebra of vector fields , in which to every zero @xmath298 some matrix lie algebra @xmath299 is assigned , and linear part of field in the neighborhood of zeroes is fixed by corresponding algebra , then such algebra has to be closed according to the commutator .
the group of algebra @xmath288 would contain the groups of algebras @xmath299 inside.
the diagrammatic technique considered in the text allows us to look at a non - linear dynamic in continuous time from algebraic point of view .
in fact , the key result is the formula ( [ general_exponent ] ) for non - linear exponent : @xmath300 which develop a map from lie algebra of the vector fields to a corresponding group .
+ another key result is an algorithm of the resonance point neighborhood dynamic calculations , described in the section perturbations theory : resonance case , and the idea of the arising of additional degrees of freedom as a consequence of the resonances . +
a scale of the considered phenomenons , in some sense , is intermediate between local , in which systems behavior is determined by constant and linear terms , and global one , in which a global geometry of vector field , singularities , zeroes mutual disposition , etc .
is sufficient .
+ in conclusion , let s consider some questions , which look worth to be discussed later : + * applications : * 1 . _ chaos theory .
_ it is well known , that chaotic dynamic could be observed in spatial dimension 3 or higher only . in context of resonances in the perturbations theory , we could suppose , that it is connected to very large number of solutions of the equations @xmath301 for particular @xmath91 values .
how bifurcations are related to the dependency of a number of solutions on parameters ? 2 .
_ hamiltonian dynamic .
_ under such consideration , in hamiltonian systems a non - linearity looks like @xmath302 , where @xmath303 - symplectic form
. an interesting question is what specific properties hamiltonian dynamic has in this context , how this applies to kam - theory ?
_ integrable systems .
_ the symmetries groups , which arise in integrable systems , are often generated by some semi - simple lie algebras .
we could suppose , that when the system leaves a regime , where system is linear up to deformation , a general group , generated by the arbitrary vector fields would play a substantial role . in its study , the examples , which arise from the polynomial vector fields could be useful .
* calculations : * 1 . _
fixed point_. how can we get the series near a fixed point not from the ordered integration , but from the general formula for evolution operator ? 2
. _ resonance phase space structure . _
how could we describe better topology of a phase space near the resonance fixed point ?
3 . _ convergence .
_ how can we compute a convergence domain of an arbitrary evolution operator , around some point ? and in perturbation serieses case ? 4 . _
jordan form .
_ how should we change ansatz , to consider dynamic around the fixed points , with non - diagonisable linear part of field ? 5 . _
_ all calculations in this text was done over field @xmath304 .
when we do calculations over @xmath165 , some linear operators could nt be completely diagonalized - some non - jordan cells is left sometimes .
what new phenomenons would it produce ? 6 . _
diagrams reduction_. how could we simplify the calculations , in case if non - linearity could be represented as a product @xmath305 ?
* conceptual : * 1 . _ approximations . _
if we approximate some function in the r.h.s of equation by polynomial of some degree , how an increase of order of this polynomial change topology of a phase space of the dynamical systems ? 2 . _
branching algebras .
_ what is the difference of the branching algebras from the ordinal one ?
how their representations theory in tensor algebra looks like ? what is the specific properties of corresponding lie groups ? 3 .
_ normal forms .
_ a linear operators classification is based on their normal form .
it is also useful in the straightforward matrix exponent calculations .
how could we classify the non - linear operators ? this way , we should get better understanding of the resultants theory ( see @xcite , @xcite ) .
_ infinite dimensional operators .
_ what results this technique could develop in case of pde ?
in fact , all which we should change is to replace a finite dimensional vector space , with the hilbert space of functions over some manifold .
another approach is a space discritization , and a continuum limit .
for example , kdv equation is a bright example of decomposition around fixed point - it could be represented as a sum of a linear operator and a polynomial non - linearity .
resonance terms for non - linearity of the second degree , and corresponding frequencies relation : + 1 . for @xmath306
@xmath307 2 .
@xmath308 2 . for @xmath309
@xmath307 ( 2 resonance diagrams ) 2 .
@xmath310 ( 2 resonance diagrams ) 3 .
@xmath311 4 .
@xmath312 5 .
@xmath313 the resonance terms for non - linearity of the third degree , and corresponding frequencies relation : 1 . for @xmath306
@xmath314 2 . for @xmath309
, there is no resonances 3 . for @xmath315 :
1 . @xmath316 2 .
@xmath317 4 . for @xmath318
@xmath316 2 .
@xmath317 5 . for @xmath319
@xmath320 ( 2 resonance diagrams ) 2 .
@xmath321 3 .
@xmath322 ( 2 resonance diagrams ) 4 .
@xmath323 6 . for @xmath324 , no resonances .
a theoretical end numerical check for the resonances have been done separately .
i am highly grateful to valerii dolotin and alexey morozov , who involved me into this research field , and for our conversation , which guided me through this work .
also i want to thank yaroslav gerasimenko , adel salakh and pavlo havrilenko for our fruitful discussions . +
this work was supported by the joint ukrainian - russian sffr - rfbr project f53.2/028 .
h. poincare .
_ les metodes nouvelles de la mecanique celeste _ , paris , 1972 v.dolotin , a.morozov . _
introduction to non - linear algebra _ , arxiv : hep - th/0609022v4 c.siegel .
_ ober die normalform analyticher differentialgleichungen in der nahe einer gleichgewichtslosung _
gottingen , math .- phys .
iia , jahrg .
21 - 30 , 1952 d.grobman .
_ topological equivalence in general systems of differential equations _ , matematicheskii sbornik , 73(115):4 , 600 - 609 , 1967 v. armold .
_ small denominators and stability problems in the celestial mechanics _ , uspekhi matematicheskikh nauk , 5(113 ) , 13 - 40 , 1963 y. manin , a. panchishkin
. _ introduction to modern numbers theory _
i.m gelfand , m.m kapranov , a.v zelevinsky , _ discriminants of polynomials in many variables _ , functional analysis and its applications , 24:1 , 14 , 1990 a.morozov , sh.shakirov .
_ new and old results in resultants theory _ , teoreticheskaya i matematicheskaya fizika , 163:2 , 222 - 257 , 2010 | in this paper we investigate phase flows over @xmath0 and @xmath1 generated by vector fields @xmath2 where @xmath3 are finite degree polynomials . with the convenient diagrammatic technique we get expressions for evolution operators @xmath4 through the series in powers of @xmath5 and @xmath6 , represented as sum over all trees of particular type .
estimates are made for the radius of convergence in some particular cases .
the phase flows behavior in the neighborhood of vector field fixed points are examined .
resonance cases are considered separately . |
consider the space @xmath2 of @xmath3 hermitian matrices of the form : @xmath4 due to the hermicity , the diagonal entries in ( [ eq : xmatrix7 ] ) are real numbers , while elements of the minor diagonal are pairwise complex conjugate numbers , @xmath5 and @xmath6 supposing that the matrix @xmath7 is semi - positive definite , @xmath8 and has a unit trace , @xmath9 the @xmath7 can be regarded as the density matrix of a 4-level quantum system . since non - zero elements in ( [ eq : xmatrix7 ] ) are distributed in a shape similar to the latin letter `` x '' , the corresponding quantum states are named as @xmath0states .
the 7-dimensional space @xmath10 represents a subspace of the 15-dimensional state space @xmath11 of a generic 4-level quantum system , @xmath12 since the introduction of @xmath0states @xcite , various subfamilies of @xmath10 have been attracting a special attention .
there are at least two reasons for that interest .
first of all , it was found that microscopic systems , being in certain @xmath13states , show a highly non - trivial quantum behaviour.states .
for further references on @xmath0states cf .
@xcite , @xcite . ]
secondly , due to the simple algebraic structure of @xmath0states , many computational difficulties , common for generic states , can be resolved dealing with this special subclass of states.states possess .
namely , it can be easily verified that every @xmath0state ( [ eq : xmatrix7 ] ) is equivalent to a block - diagonal matrix @xmath14.\ ] ] ] the aforementioned simplification turned out to be very important in describing such a complicated phenomenon as the entanglement in composite quantum systems .
particularly , it is well - known that the famous entanglement measure - concurrence - can be reduced to a simple analytical expression for @xmath0states . in the present note we will move towards a detailed entanglement classification of the mixed 2-qubit @xmath0states .
namely , the parametrization of separable mixed @xmath1states of two qubits with an arbitrary spectrum of the density matrix will be described .
our analysis in the subsequent sections includes the following steps : 1 .
two unitary groups , both acting adjointly on the 7-dimensional space of 2-qubit @xmath0states , will be introduced ; 1 .
the first one is the so - called _
`` global group '' _ , @xmath15 defined as the invariance group of the subspace @xmath16 @xmath17 2 .
the second one is the subgroup of @xmath18 the so - called _ `` local group '' _ , @xmath19 its elements have a tensor product form corresponding to the decomposition of the state space @xmath2 into two qubits subspaces , @xmath20 2 .
the `` global orbits '' , @xmath21 , of the group @xmath22 will be identified and classified into families / types according to the degeneracies in the spectrum of density matrices .
3 . considering the equivalence classes induced by the local group @xmath23 action on @xmath24 one can divide the latter into different subfamilies according to their entanglement characteristics . having in mind
this ranging , the separable density @xmath25-matrices will be categorized within the global orbits classification .
in order to prove the properties of 2-qubit @xmath0states announced above , let us start with few definitions .
@xmath26 * invariance subalgebra of @xmath0states @xmath26 * the basis for the @xmath27 algebra is constructed as follows : let @xmath28s denote the set of @xmath29 matrices , where @xmath30 is a unit matrix and @xmath31 are three pauli matrices @xmath32 the set of all possible tensor products of two copies of matrices @xmath33 , @xmath34 forms the basis of the algebra @xmath35 for our aims it is useful to write the latter as the direct sum , @xmath36 where the 6-dimensional vector space @xmath37 is composed as @xmath38 while the 9-dimensional space @xmath39 is and @xmath39 are such that @xmath40 \subset \mathfrak{l}\ , , \qquad [ \mathfrak{p } , \mathfrak{l } ] \subset \mathfrak{p}\ , , \qquad [ \mathfrak{p } , \mathfrak{p } ] \subset \mathfrak{l}\,,\ ] ] the direct sum @xmath41 is nothing else than the cartan decomposition of @xmath42 . ]
@xmath43 from now , to denote the matrices in ( [ eq;lspan ] ) and ( [ eq;pspan ] ) , the notations @xmath44 where @xmath45 runs from 1 to 15 , will be used @xmath46 @xmath0states ( [ eq : xmatrix7 ] ) expand over the subset @xmath47 of the introduced @xmath48 basis : @xmath49 the real coefficients @xmath50 in ( [ eq : xmatrexp ] ) are given by the linear combinations of the density matrix elements : @xmath51 the subset @xmath52 possesses the following properties : * the subset is closed under the matrix commutator operation , i.e. , its elements span the subalgebra of @xmath48 ; * from the commutators collected in the table 1 .
it follows that the element @xmath53 commutes with all other elements of @xmath52 ; * the remaining six elements , @xmath54 , span the @xmath55 . to check the last property
, one can construct the following linear combinations : @xmath56 and verify that their commutator relations read @xmath57=\pm 2 s_\pm\ , , \qquad [ s_+ , s_-]= 4 s_z\ , , \\ & & [ t_z , t_\pm]=\pm 2 t_\pm\ , , \qquad [ t_+ , t_-]= 4 t_z\,.\end{aligned}\ ] ] thus , two sets of elements @xmath58 generate two copies of @xmath59 algebra . gathering all together , we conclude that the set @xmath52 generates the subalgebra @xmath60-states see @xcite . ]
@xmath26 * global unitary group of @xmath0states @xmath26 * exponentiation of the algebra @xmath61 results in the 7-parametric subgroup of @xmath62 , @xmath63 whose action preserves the @xmath0states space @xmath2 , i.e. , @xmath64 using the expansion @xmath65 over the 7-tuple @xmath66 and the formulae ( [ eq : chb1])-([eq : chb2 ] ) from the section * [ sec : supmat ] .
supplementary material * , one can verify that the group @xmath67 has the following representation : @xmath68 where the two copies of su(2 ) are parametrized as follows : @xmath69}\,,\\ su(2)^\prime&= & \exp{\left[i\left(-\omega_4+\omega_7\right)\sigma_1 + i\left(-\omega_2+\omega_5\right)\sigma_2 + i\left(\omega_3 -\omega_6\right)\sigma_3 \right ] } \,.\end{aligned}\ ] ] @xmath26 * local subgroup of @xmath70 @xmath26
* suppose now that our 4-level system is composed of 2-level subsystems , i.e. , two qubits . in this case
the hilbert space @xmath71 is given by the tensor product of 2-dimensional hilbert spaces , @xmath72 and one can consider the tensor product of operators acting independently on the subspaces of individual qubits , @xmath73 and @xmath74 particularly , having in mind an intuitive idea of mutual independence of isolated qubits , we define the _ `` local unitary group '' _ , @xmath75 , as the subgroup of global invariance group of @xmath0 states , @xmath76 such that each of its elements @xmath77 has the tensor product form , @xmath78 with @xmath79 from the expression ( [ eq : g_xrepr ] ) it follows that the local unitary group can be written as : @xmath80
now it will be shown that every @xmath0states density matrix can be diagonalized using some subgroup of global @xmath82 group .
therefore , the adjoint @xmath83orbits structure is completely determined by the coset @xmath84 , where @xmath85 stands for the isotropy group of a density matrix @xmath86 .
this isotropy group , in turn , depends on the degeneracies occurring in the spectrum of density matrices .
thus , the latter determines all possible types of @xmath83orbits and the corresponding classification can be carried as follows .
consider the adjoint action of the global unitary group @xmath87 on the 7-dimensional space @xmath10 and introduce the following vectors at each point @xmath88 : @xmath89\ , , \quad k= 3 , 6 , 7 , 8 , { 10 } , { 11 } , { 15}\,.\ ] ] in the equation ( [ eq : tangvec ] ) the group elements @xmath90 are parametrized by 7-tuple @xmath91 : @xmath92 these vectors belong to a tangent space of @xmath87-orbits .
the dimensionality of this tangent space is given by the rank of the @xmath93 gram matrix @xmath94 straightforward evaluation of the spectrum @xmath95 of the gram matrix @xmath96 shows that it comprises : two eigenvalues of multiplicity 2 and three identically vanishing eigenvalues , @xmath97 where the double multiplicity eigenvalues read : @xmath98 the formulae ( [ eq : dmeigen1 ] ) and ( [ eq : dmeigen2 ] ) ensure that there exist 4 types of @xmath87-orbits : * * dim @xmath99 = 4 * , the generic orbits ; * * dim @xmath99 = 2 * , the degenerate orbits defined by the equations : @xmath100 * * dim @xmath99 = 2 * , the degenerate orbits defined by the equations : @xmath101 * * dim @xmath99 = 0 * , the single orbit @xmath102 - the maximally mixed state . in terms of the eigenvalues of density matrices ,
the 4d orbits are consistent with a generic spectrum , i.e. , matrices with 4 different eigenvalues , while 2d orbits are generated by @xmath25-matrices with double multiplicity of the following types : @xmath103 here a detailed representation for each type of @xmath83orbits will be given , starting from the orbit of the highest dimensionality .
let us assume that the spectrum of @xmath105 is a generic one , i.e. , all eigenvalues @xmath106 are different positive real numbers .
furthermore , in the block - diagonal representation ( [ eq : xmatrixblockdiag ] ) of the density matrix @xmath105 , the @xmath107 denote the eigenvalues of the upper block and @xmath108 are eigenvalues of the lower block .
the @xmath109 density matrix @xmath105 can be diagonalized in a blockwise way , @xmath110 using a special unitary matrix @xmath111 where @xmath112 and @xmath113 are @xmath29 special unitary matrices diagonalizing the upper and lower sub - blocks in ( [ eq : xmatrixblockdiag ] ) . since a generic spectrum has been assumed , matrices @xmath112 and @xmath113 belong to the coset , @xmath114 , where the group @xmath115 interchanges eigenvalues inside the pairs @xmath116 and @xmath117 . in order to have uniqueness in ( [ eq : ddiagxmatr ] ) , one can fix a certain order in the spectrum @xmath118 namely , we assume that elements of the spectrum form a partially ordered simplex , @xmath119 , i.e. , @xmath120 depicted in the figure [ fig : partorderedsimplex ] . and
satisfies the inequalities @xmath121 and @xmath122 as the image of the partially ordered simplex @xmath119 , while the tetrahedron @xmath123 inside it corresponds to a 3d simplex with the following complete order of eigenvalues : @xmath124 . ]
[ fig : partorderedsimplex ] comparing expression ( [ eq : unitdiag ] ) with ( [ eq : g_xrepr ] ) , we convinced that the diagonalizing matrix is an element of the global group @xmath125 with @xmath126 special unitary matrices @xmath112 and @xmath113 from the coset @xmath127 parametrized by angles @xmath128\ , , \psi_1,\psi_2 \in [ 0 , 2\pi]$ ] : @xmath129 the 3-dimensional isotropy group @xmath130 of generic orbits is @xmath131 this is in accordance with the maximal dimension of the @xmath87-orbits : @xmath132 summarising , the adjoint action of the global group @xmath133 determines the generic orbits , which are locally given by product of 2-spheres , @xmath134 . according to the representation ( [ eq : xmatrixdegenerat ] ) ,
two types of 2d degenerate @xmath87-orbits are generated by the matrices with degenerate @xmath29 sub - blocks , either upper or lower blocks . in the first case the isotropy group @xmath136 reads @xmath137 while for the second case @xmath136 is @xmath138 in both cases , @xmath139 and the dimension of these degenerate @xmath87-orbits is @xmath140 finally , there is one point in the state space @xmath10 , whose isotropy group coincides with the invariance group @xmath142 this point corresponds to the maximally mixed state , @xmath143
now we are in position to prove that every type of @xmath87-orbits includes the separable states . describing the mixed state of a composed system
@xmath144 is _ separable _ if it allows the convex decomposition : @xmath145 where @xmath146 and @xmath147 represent the density matrices acting on the multipliers @xmath73 and @xmath148 correspondingly .
otherwise , it is _ entangled _ @xcite . ]
[ fig : abssepx ] the separability of states as a function of density matrices spectrum @xmath149 , can be analysed using the representation ( [ eq : ddiagxmatr ] ) for the generic @xmath87-orbits . according to the peres - horodecki criterion @xcite , which is a necessary and sufficient condition for separability of @xmath29 and @xmath150 dimensional systems , a state @xmath86 is separable if its partial transposition , i.e. , @xmath151 is semi - positive as well . in the second subsystem
; similarly , one can use the alternative action , @xmath152 straightforward computation with @xmath153 in the form ( [ eq : ddiagxmatr ] ) shows that the semi - positivity of the partially transposed matrix @xmath154 requires fulfilment of the following inequalities : @xmath155 note that the inequalities ( [ eq:1postreig ] ) and ( [ eq:2postreig ] ) do not constraint two angles @xmath156 and @xmath157 in ( [ eq : uv ] ) that parametrize the local group @xmath158 it conforms with a general observation that the separability property is independent from the local characteristics of the composite system .
this local group is the factor of the global group @xmath159 , and the corresponding factor in the matrix @xmath160 diagonalising @xmath105 , is irrelevant for the separability of @xmath25- states .
analysing the inequalities ( [ eq:1postreig ] ) and ( [ eq:2postreig ] ) , one can conclude : * there are separable states for any values of eigenvalues from the partially ordered simplex @xmath119 . in other words , the inequalities ( [ eq:1postreig ] ) and ( [ eq:2postreig ] )
determine non - empty domain of definition for angles @xmath161 and @xmath162 in ( [ eq : uv ] ) for every non - degenerate spectrum @xmath149 ; * there is a special family of the so - called `` absolutely separable '' @xmath25-states , such that the angles @xmath161 and @xmath162 can be arbitrary one .
the absolutely separable @xmath25-states are generated by subset of the partially ordered simplex ( [ eq : semi - orderedsim ] ) defined by the inequalities @xmath163 the figure [ fig : abssepx ] . illustrates location of the subset of the absolutely separable states inside the partially ordered simplex @xmath164 testing the degenerate density matrices of the form ( [ eq : xmatrixdegenerat ] ) by the peres - horodecki criterion , we reveal the following picture . the positivity requirement of partially transposed density matrix with double multiplicity of eigenvalues gives the inequalities similar to ( [ eq:1postreig ] ) and ( [ eq:2postreig ] ) .
however , owing to the larger isotropy group @xmath136 of states , the new inequalities depend solely on a single coordinate of the coset @xmath165 more precisely , if @xmath166 i.e. , the degeneracy occurs in the upper sub - block , then the angle @xmath162 that parametrizes the matrix @xmath113 in ( [ eq : uv ] ) plays the role of such a coordinate . in this case
, the peres - horodecki criterion asserts that the degenerate @xmath0state is separable iff : @xmath167 where @xmath168 .
this inequality points out the critical value @xmath169 , such that for @xmath170 the angle @xmath162 is constrained , while for the interval @xmath171 the state is separable for an arbitrary angle @xmath172 the analogous results for the angle @xmath161 ( see the matrix @xmath112 in ( [ eq : uv ] ) ) hold true if the lower sub - block in ( [ eq : xmatrixdegenerat ] ) is degenerate , i.e. , @xmath173 therefore , in both classes of the degenerate 2d global orbits one can point out 2d family of separable degenerate states . furthermore ,
among them there are the `` degenerate absolutely separable '' states , i.e. , the degenerate global 2d orbits consisting completely from the separable states .
the present article is devoted to the discussion of an interplay between local and global characteristics of a pair of qubits in mixed @xmath1states . with this aim ,
orbits of the global unitary group @xmath87 action were described and classified according to the degeneracies occurring in the spectrum of density matrices . based on this analysis ,
the dependence of @xmath1states separability on the spectrum has been studied .
particularly , the separable @xmath1states have been collected into the following families : * the 4-dimensional family of separable states with the spectrum in general position ; * two classes of 2-dimensional separable states with the double degeneracy spectrum ; * the maximally mixed state .
finalizing notes , it is worth to comment that according to the aforementioned classification , the entangled states being complementary to the separable states , are partitioned likewise into three types .
however , such classification is not complete .
a further , more subtle ranging of the entangled states located on the given @xmath87-orbit into subclasses is necessary .
the latter subclasses are determined not by invariants of the global group @xmath87 , but are specified by the values of the @xmath174-invariants . in the forthcoming publications we are planning to discuss this issue in more detail .
apart from that , following the approach elaborated in @xcite , @xcite and @xcite , the generalization of the derived results for a generic case of 15-dimensional 2-qubit states will be considered .
here we collect a technical material useful for performing computations described in the main text .
it includes the basis of the lie algebra @xmath42 , commutators of its elements and block - diagonal representation for the subalgebra @xmath52 . 2 ting yu and j.h .
evolution from entanglement to decoherence of bipartite mixed `` x '' states . quantum information and computation , * 7 * , 459 - 468 ( 2007 ) .
nielsen and i.l .
quantum computation and quantum information .
cambridge university press ( 2011 ) .
quantum states with einstein - podolsky - rosen correlations admitting a hidden - variable model .
a 40(8 ) * 42774281 ( 1989 ) .
m. horodecki and p. horodecki .
reduction criterion of separability and limits for a class of distillation protocols .
rev . * a 59 * 42064216 ( 1999 ) .
s. ishizaka and t. hiroshima .
maximally entangled mixed states under nonlocal unitary operations in two qubits .
rev . * a 62 * 22310 ( 4pp ) ( 2000 ) .
f. verstraete , k. audenaert , t.d .
bie and b.d .
maximally entangled mixed states of two qubits .
rev . * a 64 * 012316 ( 6pp ) ( 2001 ) .
n.quesada , a.al-qasimi and d.f.v .
james , quantum properties and dynamics of x states , journal of modern optics , * 59 * , 13221329 ( 2012 ) .
p. mendonca , m. marchiolli and d. galetti , entanglement universality of two - qubit @xmath0states .
annals of physics . *
351 * , 79103 ( 2014 ) .
a.r.p.rau , manipulating two - spin coherences and qubit pairs , phys . rev . * a 61 * 032301 ( 2000 ) .
a.r.p.rau , algebraic characterization of x - states in quantum information journal of physics a : mathematical and theoretical , * 42 * , 412002 ( 7pp ) ( 2009 ) a.peres , separability criterion for density matrices , phys . rev .
lett . * 77 * 1413 - 1415 ( 1996 ) .
v. gerdt , a. khvedelidze and yu .
palii , on the ring of local polynomial invariants for a pair of entangled qubits , zapiski pomi * 373 * , 104 - 123 ( 2009 ) .
v. gerdt , a. khvedelidze and yu .
palii , constraints on su(2)@xmath185su(2 ) invariant polynomials for entangled qubit pair , phys . of atomic nucleii ,
74 , 893 - 900 ( 2011 ) .
v. gerdt , a. khvedelidze and yu .
palii , on the ring of local unitary invariants for mixed x - states of two qubits , arxiv:1603.03262 [ quant - ph ] . | the separable mixed 2-qubit @xmath0states are classified in accordance with degeneracies in the spectrum of density matrices .
it is shown that there are four classes of separable @xmath1states , among them : one 4d family , a pair of 2d family and a single , zero - dimensional maximally mixed state . |
since almost a decade , observations from cosmic microwave background ( cmb hereafter ) satellite , balloon - borne and ground based experiments ( @xcite , @xcite , @xcite , @xcite ) , galaxy redshift surveys @xcite and luminosity distance measurements , are fully confirming the theoretical predictions of the standard @xmath14cdm cosmological model .
this not only permits to place stringent constraints on the parameters of the model but can be fruitfully used to constrain non standard physics at the fundamental level , such as classes of elementary particle models predicting a different radiation content in the universe .
one of the major theoretical predictions of the standard scenario is the existence of a relativistic energy component ( see e.g. @xcite ) , beside cmb photons , with a current energy density given by : @xmath15\rho{_\gamma } \ , \ ] ] where @xmath16 is the energy density of the cmb photons background at temperature @xmath17 and @xmath0 is in principle a free parameter , defined as the effective number of relativistic degrees of freedom . assuming standard electroweak interactions , three active massless neutrinos and including the ( small ) effect of neutrino flavour oscillations the expected value is @xmath18 with a deviation from @xmath19 that takes into account effects from the non - instantaneous neutrino decoupling from the primordial photon - baryon plasma ( see e.g. @xcite ) . in recent years , thanks to the continuous experimental advancements , the value of @xmath0 has been increasingly constrained from cosmology ( @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) , ruling out @xmath20 at high significance .
however , especially after the new act @xcite and spt @xcite cmb results , the data seem to suggest values higher than the `` standard '' one , with @xmath21 ( see e.g. @xcite , @xcite , @xcite , @xcite , @xcite ) in tension with the expected standard value at about two standard deviations .
the number of relativistic degrees of freedom obviously depends on the decoupling process of the neutrino background from the primordial plasma .
however , a value of @xmath22 is difficult to explain in the three neutrino framework since non - standard neutrino decoupling is expected to maximally increase this value up to @xmath23 ( see e.g. @xcite ) .
a possible explanation could be the existence of a fourth ( or fifth ) sterile neutrino .
the hypothesis of extra neutrino flavour is interesting since recent results from short - baseline neutrino oscillation data from lsnd @xcite and miniboone @xcite experiments are consistent with a possible fourth ( or fifth ) sterile neutrino specie ( see @xcite and references therein ) .
moreover , a larger value for @xmath24 could arise from a completely different physics , related to axions ( see e.g. @xcite ) , gravity waves ( @xcite ) , decaying particles ( see e.g. @xcite ) , extra dimensions @xcite and dark energy ( see e.g. @xcite and references therein ) . as a matter of fact ,
any physical mechanism able to produce extra `` dark '' radiation produces the same effects on the background expansion of additional neutrinos , yielding a larger value for @xmath0 from observations . since there is a large number of models that could enhance @xmath0 it is clearly important to investigate the possible ways to discriminate among them .
if dark radiation is made of relativistic particles as sterile neutrinos it should behave as neutrinos also from the point of view of perturbation theory , i.e. if we consider the set of equations that describes perturbations in massless neutrino ( following the definition presented in @xcite ) : @xmath25 it should have an effective sound speed @xmath3 and a viscosity speed @xmath4 such that @xmath26 .
free streaming of relativistic neutrino will indeed produce anisotropies in the neutrino background yielding a value of @xmath27 while a smaller value would indicate possible non standard interactions ( see e.g. @xcite ) .
a value of @xmath4 different from zero , as expected in the standard scenario , has been detected in @xcite and confirmed in subsequent papers @xcite .
more recently , the analysis of @xcite confirmed the presence of anisotropies from current cosmological data but also suggested the presence of a lower value for the effective sound speed with @xmath28 ruled out at more than two standard deviations .
given the current situation and the experimental hints for @xmath29 is therefore timely to perform a new analysis for @xmath0 ( and the perturbation parameters @xmath7 and @xmath8 ) with the most recent cosmological data .
this is the kind of analysis we perform in this paper , organizing our work as follows : in sec .
ii we describe the data and the data analysis method adopted .
we present our results in the first two subsections of sec .
iii , depending on two adopted different parametrizations for the dark radiation . moreover a model independent analysis is also discussed in the last subsection of sec .
finally we conclude in sec . iv .
we perform a cosmomc @xcite analysis combining the following cmb datasets : wmap7 @xcite , acbar @xcite , act @xcite , and spt @xcite , and we analyze datasets using out to @xmath30 .
we also include information on dark matter clustering from the galaxy power spectrum extracted from the sdss - dr7 luminous red galaxy sample @xcite .
finally , we impose a prior on the hubble parameter based on the last hubble space telescope observations @xcite .
the analysis method we adopt is based on the publicly available monte carlo markov chain package ` cosmomc ` @xcite with a convergence diagnostic done through the gelman and rubin statistic .
we sample the following six - dimensional standard set of cosmological parameters , adopting flat priors on them : the baryon and cold dark matter densities @xmath31 and @xmath32 , the ratio of the sound horizon to the angular diameter distance at decoupling @xmath33 , the optical depth to reionization @xmath34 , the scalar spectral index @xmath35 , and the overall normalization of the spectrum @xmath36 at @xmath37 .
we consider purely adiabatic initial conditions and we impose spatial flatness .
we vary the effective number of relativistic degrees of freedom @xmath0 , the effective sound speed @xmath7 , and the viscosity parameter @xmath8 . in some cases , we consider only variations in the _ extra _ dark radiation component @xmath38 , varying the perturbation parameters @xmath4 and @xmath3 only for this extra component and assuming @xmath26 for the standard @xmath10 neutrino component
. in our analysis we always fix the primordial helium abundance to the observed value @xmath39 .
this procedure is different from the one adopted , for example , in @xcite , where the @xmath40 parameter is varied assuming big bang nucleosynthesis for each values of @xmath0 and @xmath31 in the chain . since the cosmological epoch and the energy scales probed by bbn are dramatically different from the ones probed by cmb and large scale structure we prefer to do not assume standard bbn in our analysis and to leave the primordial helium abundance as fixed to a value consistent with current observations
. we account for foregrounds contributions including three extra amplitudes : the sz amplitude @xmath41 , the amplitude of clustered point sources @xmath42 , and the amplitude of poisson distributed point sources @xmath43 .
we marginalize the contribution from point sources only for the act and spt data , based on the templates provided by @xcite .
we quote only one joint amplitude parameter for each component ( clustered and poisson distributed ) . instead
, the sz amplitude is obtained fitting the wmap data with the wmap own template , while for spt and act it is calculated using the @xcite sz template at 148 ghz .
again , this is different from the analysis performed in @xcite where no sz contribution was considered for the wmap data .
as stated in the previous section , we perform two different analyses . in the first analysis we vary the amplitude of the whole relativistic contribution changing @xmath0 and the corresponding perturbation parameters @xmath8 and @xmath7 . in the second analysis
we assume the existence of a standard neutrino background and vary only the extra component @xmath38 considering only in this extra component the variations in @xmath8 and @xmath7 . in table [ standard ]
we report the constraints on the cosmological parameters varying @xmath0 with and without variations in perturbation theory .
we consider two cases : first we run our analysis fixing the perturbation parameters to the standard values , i.e. @xmath26 , then we let those parameters to vary freely .
.mcmc estimation of the cosmological parameters assuming @xmath0 relativistic neutrinos .
upper bounds at @xmath2 c.l . are reported for foregrounds parameters .
we quote the one - dimensional marginalized @xmath44 and @xmath45 c.l . for the neutrino parameters . [ cols="<,^,^ " , ] as we can see , when masses in the active neutrinos are considered , there is a slightly stronger evidence for the extra background with @xmath46 .
this is can be explained by the degeneracy present between @xmath47 and @xmath48 , well known in the literature ( see e.g. @xcite ) and clearly shown in figure [ sum - nus ] where we report the 2d marginalized contours in the plane @xmath49 . at 68% and 95% c.l . .
] for @xmath0 .
the dashed lines represent the @xmath50 and @xmath2 c.l . for a gaussian likelihood ( @xmath51 and @xmath52 ) respectively . ]
recently , in @xcite , a model - independent analysis for the extra relativistic degrees of freedom in cosmological data has been performed claiming no statistically significant evidence for it .
this simple analysis consists in extracting the maximum likelihood value @xmath53 as a function of @xmath0 over the parameter space sampled in the chains , with a bin width of @xmath54 and constructing a profile likelihood ratio by considering @xmath55 as a function of @xmath0 ; where @xmath56 is the maximum likelihood in the entire chains .
here we perform a similar analysis , using however a smaller bin width of @xmath57 and considering the case where the whole number of relativistic degrees of freedom @xmath0 is varied while @xmath58 .
the resulting likelihood ratio @xmath59 , plotted in figure [ maxlike ] , clearly indicates a preference for a dark radiation component finding that the best fit model has @xmath60 with a @xmath61 respect to the best fit model with @xmath18 .
we should however point out that the ratio @xmath59 presented in figure [ maxlike ] is rather noisy .
bayesian methods such as mcmc are indeed known to be inaccurate for this purpose ( see for example the discussion in @xcite ) .
other methods more appropriate for a frequentist analysis have been presented , for example , in @xcite .
in this paper we performed a new search for dark radiation , parametrizing it with an effective number of relativistic degrees of freedom @xmath0 .
we have shown that the cosmological data we considered are clearly suggesting the presence for an extra dark radiation component with @xmath1 at @xmath2 c.l . .
performing an analysis on its effective sound speed @xmath3 and viscosity @xmath4 parameters , we found @xmath5 and @xmath6 at @xmath2 c.l . ,
consistent with the expectations of a relativistic free streaming component ( @xmath7=@xmath8=@xmath9 ) .
assuming the presence of @xmath10 standard relativistic neutrinos we constrain the extra dark radiation component with @xmath11 and @xmath12 at @xmath2 c.l . while @xmath8 is practically unconstrained .
assuming a mass in the @xmath10 neutrino component we obtain further indications for the dark radiation component with @xmath13 at @xmath2 c.l . .
from these results we conclude that dark radiation currently represents one of the most relevant anomaly for the @xmath14-cdm scenario
. when comparison is possible , our results are in good agreement with the most recent analysis presented in @xcite that uses a different choice of datasets ( for example , we do nt consider matter fluctuations data from lyman-@xmath62 as in @xcite ) and an independent analysis method .
dark radiation will be severely constrained in the very near future by the planck satellite data , where a precision on @xmath0 of about @xmath63 is expected ( see e.g. @xcite and @xcite ) only from cmb data .
we thank ryan keisler for providing us with the likelihood code for the spt data .
we thank luca pagano for help . this work is supported by prin - inaf , `` astronomy probes fundamental physics '' .
support was given by the italian space agency through the asi contracts euclid- ic ( i/031/10/0 ) .
r. keisler _ et al .
_ , arxiv:1105.3182 [ astro-ph.co ] . b. a. reid _ et al .
_ , mon . not .
* 404 * ( 2010 ) 60 [ arxiv:0907.1659 [ astro-ph.co ] ] .
e. w. kolb and m. s. turner , front .
* 69 * ( 1990 ) 1 .
g. mangano , g. miele , s. pastor , t. pinto , o. pisanti , p. d. serpico , nucl .
* b729 * ( 2005 ) 221 - 234 .
[ hep - ph/0506164 ] .
a. aguilar _ et al . _
[ lsnd collaboration ] , phys .
d * 64 * ( 2001 ) 112007 [ arxiv : hep - ex/0104049 ] .
a. a. aguilar - arevalo _ et al . _
[ the miniboone collaboration ] , phys .
* 98 * ( 2007 ) 231801 [ arxiv:0704.1500 [ hep - ex ] ] ; a. a. aguilar - arevalo _ et al . _
[ miniboone collaboration ] , phys .
* 103 * ( 2009 ) 111801 [ arxiv:0904.1958 [ hep - ex ] ] .
s. hannestad , a. mirizzi , g. g. raffelt and y. y. y. wong , jcap * 1008 * ( 2010 ) 001 [ arxiv:1004.0695 [ astro-ph.co ] ] ; a. melchiorri , o. mena and a. slosar , phys .
d * 76 * ( 2007 ) 041303 [ arxiv:0705.2695 [ astro - ph ] ] .
k. nakayama , f. takahashi and t. t. yanagida , phys .
b * 697 * ( 2011 ) 275 [ arxiv:1010.5693 [ hep - ph ] ] .
t. l. smith , e. pierpaoli and m. kamionkowski , phys .
rev . lett .
* 97 * ( 2006 ) 021301 [ arxiv : astro - ph/0603144 ] .
w. fischler , j. meyers , phys .
* d83 * ( 2011 ) 063520 .
[ arxiv:1011.3501 [ astro-ph.co ] ] ; k. ichikawa , m. kawasaki , k. nakayama , m. senami , f. takahashi , `` increasing effective number of neutrinos by decaying particles , '' jcap * 0705 * ( 2007 ) 008 .
[ hep - ph/0703034 [ hep - ph ] ] ; k. nakayama , f. takahashi , t. t. yanagida , `` a theory of extra radiation in the universe , '' phys .
* b697 * , 275 - 279 ( 2011 ) .
[ arxiv:1010.5693 [ hep - ph ] ] p. binetruy , c. deffayet , u. ellwanger and d. langlois , phys .
b * 477 * ( 2000 ) 285 [ arxiv : hep - th/9910219 ] ; t. shiromizu , k. i. maeda and m. sasaki , phys .
d * 62 * ( 2000 ) 024012 [ arxiv : gr - qc/9910076 ] ; v. v. flambaum and e. v. shuryak , europhys . lett . * 74 * ( 2006 ) 813 [ arxiv : hep - th/0512038 ] .
a. hebecker and j. march - russell , nucl .
b * 608 * ( 2001 ) 375 [ arxiv : hep - ph/0103214 ] .
e. calabrese , d. huterer , e. v. linder , a. melchiorri and l. pagano , phys .
d * 83 * , 123504 ( 2011 ) [ arxiv:1103.4132 [ astro-ph.co ] ] ; j. s. gagnon and j. lesgourgues , jcap * 1109 * ( 2011 ) 026 [ arxiv:1107.1503 [ astro-ph.co ] ] .
w. hu , astrophys . j. * 506 * ( 1998 ) 485 - 494 .
[ astro - ph/9801234 ] ; w. hu , d. j. eisenstein , m. tegmark , m. j. white , phys . rev . *
d59 * , 023512 ( 1999 ) .
[ astro - ph/9806362 ] .
j. f. beacom , n. f. bell , s. dodelson , phys .
* 93 * ( 2004 ) 121302 .
[ astro - ph/0404585 ] ; s. hannestad , jcap * 0502 * ( 2005 ) 011 .
[ astro - ph/0411475 ] ; a. basboll , o. e. bjaelde , s. hannestad and g. g. raffelt , phys .
d * 79 * ( 2009 ) 043512 [ arxiv:0806.1735 [ astro - ph ] ] ; r. trotta , a. melchiorri , phys . rev .
* 95 * ( 2005 ) 011305 .
[ astro - ph/0412066 ] .
h. trac , p. bode and j. p. ostriker , astrophys .
j. * 727 * ( 2011 ) 94 [ arxiv:1006.2828 [ astro-ph.co ] ] . f. de bernardis , a. melchiorri , l. verde and r. jimenez , jcap * 0803 * ( 2008 ) 020 [ arxiv:0707.4170 [ astro - ph ] ] .
a. vikhlinin _ et al .
_ , astrophys . j. * 692 * ( 2009 ) 1060 [ arxiv:0812.2720 [ astro - ph ] ] .
a. x. gonzalez - morales , r. poltis , b. d. sherwin and l. verde , arxiv:1106.5052 [ astro-ph.co ] . s. galli , m. martinelli , a. melchiorri , l. pagano , b. d. sherwin , d. n. spergel , phys .
* d82 * ( 2010 ) 123504 .
[ arxiv:1005.3808 [ astro-ph.co ] ] .
y. akrami , p. scott , j. edsjo , j. conrad and l. bergstrom , jhep * 1004 * , 057 ( 2010 ) [ arxiv:0910.3950 [ hep - ph ] ] ; f. feroz , k. cranmer , m. hobson , r. ruiz de austri and r. trotta , jhep * 1106 * ( 2011 ) 042 [ arxiv:1101.3296 [ hep - ph ] ] .
s. hannestad , jcap * 0305 * ( 2003 ) 004 [ astro - ph/0303076 ] ; v. barger , j. p. kneller , h. -s .
lee , d. marfatia and g. steigman , phys .
b * 566 * , 8 ( 2003 ) [ hep - ph/0305075 ] ; p. crotty , j. lesgourgues and s. pastor , phys .
d * 67 * , 123005 ( 2003 ) [ astro - ph/0302337 ] ; k. ichikawa , m. kawasaki and f. takahashi , jcap * 0705 * , 007 ( 2007 ) [ astro - ph/0611784 ] ; j. hamann , arxiv:1110.4271 [ astro-ph.co ] . | combined analyses of recent cosmological data are showing interesting hints for the presence of an extra relativistic component , coined dark radiation . here
we perform a new search for dark radiation , parametrizing it with an effective number of relativistic degrees of freedom parameter , @xmath0 .
we show that the cosmological data we considered are clearly suggesting the presence for an extra relativistic component with @xmath1 at @xmath2 c.l .. performing an analysis on dark radiation sound speed @xmath3 and viscosity @xmath4 parameters , we found @xmath5 and @xmath6 at @xmath2 c.l . ,
consistent with the expectations of a relativistic free streaming component ( @xmath7=@xmath8=@xmath9 ) .
assuming the presence of @xmath10 relativistic neutrinos we constrain the extra relativistic component with @xmath11 and @xmath12 at @xmath2 c.l . while @xmath8 results as unconstrained .
assuming a massive neutrino component we obtain further indications for dark radiation with @xmath13 at @xmath2 c.l . . |
recent experiments@xcite have shown that magnetic - field dependences of thermodynamic quantities in the vortex state of type - ii superconductors provide unique information on the pairing symmetry and gap anisotropy .
on the theoretical side , however , calculations of those quantities still remain a fairly difficult task to perform .
the quasiclassical equations derived by eilenberger@xcite provide a convenient starting point for this purpose .
pesch@xcite obtained a compact analytic solution to them based on the lowest landau - level approximation for the pair potential .
it has been used recently to discuss the field dependences of the thermal conductivity,@xcite the density of states,@xcite and thermodynamic quantities.@xcite on the other hand , klein@xcite obtained a full numerical solution for the vortex - lattice state using a standard procedure to solve ordinary differential equations .
this numerical approach has been used extensively by ichioka _
et al_.@xcite to clarify the field dependences of the pair potential and the density of states for type - ii superconductors with various energy - gap structures
. it should be noted , however , that those numerical studies all adopted simplified model fermi surfaces instead of complicated fermi surfaces for real materials .
indeed , recent theoretical studies@xcite have clarified that detailed fermi - surface structures are indispensable for the quantitative description of the vortex state . with these backgrounds ,
we here develop an alternative method to solve the eilenberger equations for the vortex - lattice states .
a key point lies in expanding the pair potential and the quasiclassical @xmath2 function in the basis functions of the vortex - lattice states , thereby transforming the differential equations into algebraic equations .
this method has been powerful for ( i ) solving the ginzburg - landau equations@xcite and the bogoliubov - de gennes equations@xcite and ( ii ) obtaining quantitative agreements on the upper critical field @xmath3 of nb , nbse@xmath4 , and mgb@xmath4 with fermi surfaces from first - principles electronic - structure calculations.@xcite thus , the method may be more advantageous for the calculations of thermodynamic quantities in finite magnetic fields when realistic fermi surfaces are used as inputs .
it will also be convenient for microscopically calculating responses of the vortex - lattice state to external perturbations,@xcite such as vortex - lattice oscillations .
this method is applied here to calculate magnetic - field dependences of the magnetization , the entropy , the pauli paramagnetism , and the pair potential at various temperatures for the two - dimensional @xmath0- and @xmath1-wave superconductors in the clean and dirty limits .
these quantities have been obtained near @xmath3 for the @xmath0-wave pairing.@xcite our purpose here is to clarify the overall field dependence of those quantities .
this paper is organized as follows .
section ii gives the formulation .
section iii presents numerical results .
section iv summarizes the paper .
we use @xmath5 throughout .
we take the external magnetic field @xmath6 along the @xmath7 axis and express the vector potential as @xcite @xmath8 here @xmath9 is the average flux density produced jointly by the external current and the internal supercurrent , and @xmath10 is the spatially varying part of the magnetic field satisfying @xmath11 .
we choose the gauge such that @xmath12 .
the eilenberger equation for the even - parity pairing without pauli paramagnetism is given by@xcite [ eilens ] @xmath13 here @xmath14 @xmath15 is the matsubara energy with @xmath16 the temperature , @xmath17 is the relaxation time by nonmagnetic impurity scattering , and @xmath18 denotes the fermi - surface average : @xmath19 with @xmath20 an infinitesimal area on the fermi surface , @xmath21 the density of states per spin and per unit volume at the fermi energy in the normal state , and @xmath22 the fermi velocity . the operator @xmath23 in eq .
( [ eilen ] ) is defined by @xmath24 with @xmath25 the flux quantum,@xcite @xmath26 is the pair potential , and @xmath27 specifies the gap anisotropy satisfying @xmath28 . finally , the quasiclassical green s functions @xmath2 and @xmath29 for @xmath30 are connected by @xmath31 with @xmath32 . equation ( [ eilen ] ) has to be solved simultaneously with the self - consistency equation for the pair potential and the maxwell equation for @xmath33 , which are given respectively by @xmath34 , \label{pair}\ ] ] @xmath35 with @xmath36 the transition temperature for @xmath37 .
finally , the free - energy functional corresponding to eq .
( [ eilens ] ) is given by@xcite @xmath38 \biggr\ } , \label{f}\end{aligned}\ ] ] where @xmath39 is the free energy in the normal state and @xmath40 is defined by @xmath41 indeed , functional differentiations of eq .
( [ f ] ) with respect to @xmath2 , @xmath42 , and @xmath33 lead to eqs.([eilen ] ) , ( [ pair ] ) , and ( [ eq : maxwell ] ) , respectively .
we first express the gradient operator in eq .
( [ eilen ] ) as @xmath43 here @xmath44 and @xmath45 are the boson operators : [ aav ] @xmath46 = \frac{l_{{\rm c}}}{\sqrt{2 } } \left [ \begin{array}{cc } c_{1 } & ic_{2 } \\ -c_{1}^ { * } & ic_{2}^ { * } \end{array } \right ] \left [ \begin{array}{c } \nabla_{x } \\
\nabla_{y}-2\pi i bx/\phi_{0 } \end{array } \right ] \
, , \label{aa*}\ ] ] with @xmath47 and @xmath48 , and @xmath49 and @xmath50 are defined by @xmath51 @xmath52 respectively .
the constants ( @xmath53,@xmath54 ) can be fixed appropriately to make the subsequent calculations efficient .
a convenient choice@xcite is to impose the condition that the gradient term in the ginzburg - landau equation be expressed in terms of @xmath55 without using @xmath56 and @xmath57 , i.e. , the pair potential near @xmath58 be described in terms of the lowest landau level only .
alternatively , one may follow graser _ et al_. @xcite to change them at every temperature and magnetic field so as to make the free - energy within the lowest - landau - level approximation smallest . using eq .
( [ aa * ] ) , we can make up a set of basis functions to describe arbitrary vortex - lattice structures as@xcite @xmath59 \nonumber \\ & & \hspace{11 mm } \times\exp\!\left [ i \frac{na_{1x}}{l_{\rm c}^{2}}\!\left(y+ l_{\rm c}^{2}q_{x}- \frac{na_{1y}}{2}\right ) \!\right ] \nonumber \\ & & \hspace{11 mm } \times \exp\!\left [ -\frac{c_{1}c_{2}}{2}\left(\ ! \frac{x-
l_{\rm c}^{2}q_{y}- na_{1x}}{c_{1}l_{\rm c}}\!\right)^{\!\ !
2 } \right ] \nonumber \\ & & \hspace{11mm}\times \frac{1}{\sqrt{2^n n!}}h_{n}\!\!\left(\ ! \frac{x- l_{\rm
c}^{2}q_{y}- na_{1x}}{c_{1}l_{\rm c}}\!\right ) \ , . \label{basis}\end{aligned}\ ] ] here @xmath60 denotes the landau level , @xmath61 is an arbitrary chosen magnetic bloch vector characterizing the broken translational symmetry of the vortex lattice and specifying the core locations , and @xmath62 is the volume of the system .
the quantities @xmath63 , @xmath64 and @xmath65 are the components of the basic vectors @xmath66 and @xmath67 in the @xmath68 plane , respectively , with @xmath69 and @xmath70 , @xmath71 denotes the number of the flux quantum in the system , and @xmath72 is the hermite polynomial .
the basis functions are both orthonormal and complete , satisfying @xmath73 and @xmath74 .
we now expand @xmath42 , @xmath2 , and @xmath49 in the basis functions of the vortex lattice as [ dfexpand ] @xmath75 @xmath76 @xmath77 where @xmath78 is the reciprocal - lattice vector.@xcite substituting eq .
( [ dfexpand ] ) into eq .
( [ eilens ] ) and using the orthogonality of the basis functions , eq .
( [ eilens ] ) is transformed into a set of algebraic equations for @xmath79 , @xmath80 , and @xmath81 as [ eilens2 ] @xmath82 @xmath83 , \label{pair2}\ ] ] @xmath84 with @xmath85 together with the equation to determine @xmath3 derived recently,@xcite the above coupled equations form a basis for efficient numerical calculations of the eilenberger equations for vortex - lattice states with arbitrary fermi - surface structures . for a given vortex - lattice structure specified by the basic vectors @xmath66 and @xmath67 in eq .
( [ basis ] ) , the coupled equation ( [ eilens2 ] ) may be solved iteratively in order of eqs .
( [ eilen2 ] ) , ( [ pair2 ] ) , and ( [ eq : maxwell2 ] ) by adopting a standard technique to solve nonlinear equations.@xcite a convenient starting point is to put @xmath86 , @xmath87 , and @xmath88 on the right - hand of eq .
( [ eilen2 ] ) , where @xmath89 is the angle - averaged energy gap in zero field . in this connection
, it may be worth noting that the tridiagonal matrix constructed from the coefficients of @xmath90 on the left - hand side of eq .
( [ eilen2 ] ) can be inverted analytically.@xcite the fermi - surface integrals @xmath91 can be performed as described in sec .
iv of ref . .
in contrast , integrations over @xmath92 in eqs.([eilen2 ] ) and ( [ eq : maxwell2 ] ) may be carried out as follows : at the beginning of each calculation , we prepare @xmath93 and @xmath94 at equally spaced @xmath95 discrete points in a unit cell .
we then construct @xmath26 , @xmath96 , @xmath97 , and @xmath31 on those points by eqs .
( [ dfexpand ] ) with restricting the summations to those satisfying @xmath98 and @xmath99 , where @xmath100 and @xmath101 are some cutoffs .
now , the integrations can be performed by the trapezoidal rule ; its convergence is excellent for periodic functions .
also , the summation over @xmath102 in eq .
( [ pair2 ] ) is restricted in the actual calculations to those satisfying @xmath103 .
the convergence can be checked by increasing @xmath104 , @xmath100 , @xmath101 , and @xmath105 .
finally , the vortex - lattice structure can be fixed by requiring that the free energy ( [ f ] ) be minimum .
once @xmath42 , @xmath2 , and @xmath33 are determined as above , we can calculate thermodynamic quantities of the vortex - lattice state . specifically , the magnetization @xmath106 due to supercurrent and the entropy @xmath107 are obtained by@xcite [ thermo ] @xmath108 @xmath109 \biggr\ } \ : , \label{eq : entropy}\end{aligned}\ ] ] respectively , where @xmath110 is the entropy in the normal state and @xmath40 is defined by eq .
( [ i ] ) . also , when it is much smaller than the diamagnetism by supercurrent , the magnetization @xmath111 due to pauli paramagnetism can be calculated by@xcite @xmath112 \ : , \label{eq : spin susceptibility}\ ] ] where @xmath113 with @xmath114 the bohr magneton . the quantity @xmath115 in eq .
( [ eq : spin susceptibility ] ) may be obtained either by numerical differentiations or directly from the equation of differentiating eq .
( [ eilen ] ) with respect to @xmath116 : @xmath117 this equation can be solved similarly as eq .
( [ eilen ] ) .
we now present numerical results for two - dimensional systems with the cylindrical fermi surface which is placed in the @xmath68 plane perpendicular to @xmath6 .
we have considered a couple of energy gaps : @xmath118 then it is convenient to set @xmath119 in eq .
( [ aav ] ) .
there is another parameter in the system corresponding to the ginzburg - landau parameter @xmath120 .
we have fixed it by @xmath121 with @xmath122 the angle - averaged energy gap at @xmath123 .
it follows from eq .
( 46 ) of ref .
that eq .
( [ kappa_0 ] ) corresponds to @xmath124 and @xmath125 for the @xmath0- and @xmath1-wave pairings in the clean limit , respectively , and @xmath126 for the @xmath0-wave pairing with @xmath127 .
the vortex - lattice structure has been fixed as hexagonal ( square ) for the @xmath0-wave ( @xmath128-wave ) pairing so that finite contributions in the expansion ( [ dexpand ] ) come only from @xmath129 ( @xmath130 ) landau levels . in the clean limit .
( a ) @xmath0-wave ; ( b ) @xmath1-wave .
the temperatures are @xmath131 , @xmath132 , @xmath133 , and @xmath134 from top to bottom in the high - field region . ]
equation ( [ eilens2 ] ) for the above model has been solved with the procedure of sec .
iid over @xmath135 .
the value @xmath136 is chosen as @xmath137 ( @xmath138 ) for the @xmath0-wave ( @xmath128-wave ) pairing .
on the other hand , we have set @xmath139 ( @xmath140 ) at @xmath141 ( @xmath142 ) .
these values have been enough to get the convergence .
using @xmath42 , @xmath2 , and @xmath49 thus obtained , we have calculated the magnetization by supercurrent , the entropy , and the magnetization by pauli paramagnetism by eqs .
( [ eq : magnetization ] ) , ( [ eq : entropy ] ) , and ( [ eq : spin susceptibility ] ) , respectively . in eq .
( [ dexpand ] ) as a function of @xmath143 in the clean limit at @xmath144 : ( a ) @xmath0-wave pairing where the curves are @xmath145 , @xmath146 , and @xmath147 from top to bottom ; ( b ) @xmath1-wave pairing where the curves are @xmath145 , @xmath148 , and @xmath149 from top to bottom near @xmath3 . ]
figure [ fig : mb - tau10000 ] presents magnetic - field dependence of the magnetization by supercurrent for the @xmath0- and @xmath1-wave pairings in the clean limit at several temperatures . in both cases ,
the initial slope at @xmath150 gradually decreases as the temperature is lowered , implying a monotonic increase of the maki parameter@xcite @xmath151 as @xmath152 .
this feature of @xmath151 has also been predicted in the case of the three - dimensional spherical fermi surface with the @xmath0-wave pairing.@xcite unlike the three dimensional case,@xcite however , the slope in these two - dimensional cases remains finite and does not approach @xmath153 even in the clean limit of @xmath154 , in agreement with a previous calculation of @xmath155.@xcite the curves at low temperatures become more and more concave upward , thereby compensating the initial reduction of the magnetization .
the temperature variation is slightly larger for the @xmath1-wave pairing than the @xmath0-wave pairing . as a function of @xmath143 .
( a ) @xmath0-wave ; ( b ) @xmath1-wave .
the temperatures are @xmath131 , @xmath132 , @xmath133 , and @xmath134 from bottom to top .
they are normalized to vary from @xmath156 at @xmath157 to @xmath153 at @xmath158 . ]
.the exponent @xmath159 of eq .
( [ exponents ] ) for the @xmath0- and @xmath1-wave pairings in the clean limit calculated by the best fit to the numerical data of fig . 3 .
[ cols="^,^,^,^,^",options="header " , ] figure [ fig : sbspinb - s - tau001 ] shows the field dependences of @xmath107 and @xmath160 for @xmath127 at various temperatures .
table iii presents the corresponding exponents @xmath161 and @xmath162 obtained from the data of @xmath163 ; unlike the clean - limit case , it has been impossible to fit the whole region by a single exponent , especially at intermediate temperatures , as may be realized from fig .
[ fig : sbspinb - s - tau001 ] . compared with figs .
3(a ) and 4(a ) , the curves are more monotonic with the almost linear behavior @xmath164 .
looking at the temperature dependence more closely , however , we observe a change from a convex - upward behavior at high temperatures to a convex - downward behavior at low temperatures , in agreement with a previous calculation near @xmath3.@xcite this feature also appears in the field dependence of the zero - energy density of states as calculated recently by miranovi _
et al._@xcite the convex - downward behavior at @xmath165 may become more pronounced at lower temperatures to be observable experimentally .
we have developed an alternative method to solve the eilenberger equations for the vortex - lattice state .
the main analytic formulas are given in sec .
iic together with the numerical procedure to solve them in sec .
this method , which directly extends the @xmath3 equation@xcite to lower fields , has a potential applicability to systems with complicated fermi surfaces and/or gap structures to carry out detailed calculations on the field dependences of thermodynamic quantities for various type - ii superconductors . using it , we have calculated the field dependences of the magnetization by supercurrent , the mixing of higher landau levels in the pair potential , the entropy , and the pauli paramagnetism for the two - dimensional @xmath0- and @xmath1-wave pairings in the clean and dirty limits at various temperatures .
previous results near @xmath3 for the @xmath0-wave pairing@xcite have been reproduced adequately and extended to lower fields to clarify the overall field dependences .
the differences between the @xmath0- and @xmath1-wave pairings are quite small at high temperatures but develop gradually as the temperature is lowered , reflecting the marked difference in the low - energy density of states between the two cases .
the field dependences of the entropy and pauli paramagnetism in the clean limit at low temperatures present convex - upward behaviors for both pairings .
in contrast , the curves of the @xmath0-wave pairing in the dirty limit are more monotonic and fairly close to the linear behavior , but also acquire downward curvature at low temperatures . as for the magnetization by supercurrent
, there is a wide region of linear field dependence from @xmath3 both at high temperatures and in the dirty limit .
the region shrinks in the clean limit as the temperature is lowered , and the curve acquires pronounced upward curvature .
it is also found that the mixing of higher landau levels in the pair potential is small for @xmath166 but develops rapidly as the field is further decreased .
this work is supported by a grant - in - aid for scientific research from the ministry of education , culture , sports , science , and technology of japan , and also from the 21st century coe program `` topological science and technology , '' hokkaido university . | we develop an alternative method to solve the eilenberger equations numerically for the vortex - lattice states of type - ii superconductors . using it , we clarify the magnetic - field and impurity - concentration dependences of the magnetization , the entropy , the pauli paramagnetism , and the mixing of higher landau levels in the pair potential for two - dimensional @xmath0- and @xmath1-wave superconductors with the cylindrical fermi surface . |
squares and square roots are classical concepts in graph theory that are defined as follows .
square _ @xmath7 of a graph @xmath8 is the graph with vertex set @xmath9 such that any two distinct vertices @xmath10 are adjacent in @xmath7 if and only if @xmath11 and @xmath12 are of distance at most 2 in @xmath13 .
a graph @xmath14 is a _ square root _ of @xmath13 if @xmath15 .
there exist graphs with no square root , graphs with a unique square root as well as graphs with many square roots .
mukhopadhyay @xcite showed in 1967 that a connected graph @xmath13 with @xmath1 vertices @xmath16 has a square root if and only if there exists a set of @xmath1 complete subgraphs @xmath17 of @xmath13 with @xmath18 such that @xmath19 contains @xmath20 for all @xmath21 , and @xmath19 contains @xmath22 if and only if @xmath23 contains @xmath20 for all @xmath24 .
this characterization did not yield a polynomial time algorithm for recognizing squares .
in fact , in 1994 , motwani and sudan @xcite showed that the square root problem , which is that of testing whether a graph has a square root , is np - complete .
this fundamental result triggered a lot of research on the computational complexity of recognizing squares of graphs and computing square roots under the presence of additional structural assumptions .
in particular , the following two recognition questions have attracted attention ; here @xmath25 denotes some fixed graph class . *
how hard is it to recognize squares of graphs of @xmath26 ? * how hard is is to recognize graphs of @xmath26 that have a square root ?
ross and harary @xcite characterized squares of a tree and proved that if a connected graph has a unique tree square root , then this root is unique up to isomorphism .
lin and skiena @xcite gave linear time algorithms for recognizing squares of trees and planar graphs with a square root .
the results for trees @xcite were generalized to block graphs by le and tuy @xcite .
lau @xcite gave a polynomial time algorithm for recognizing squares of bipartite graphs .
lau and corneil @xcite gave a polynomial time algorithm for recognizing squares of proper interval graphs and showed that the problems of recognizing squares of chordal graphs , squares of split graphs , and chordal graphs with a square root are all three np - complete .
le and tuy @xcite gave a quadratic time algorithm for recognizing squares of strongly chordal split graphs .
milanic and schaudt @xcite gave linear time algorithms for recognizing trivially perfect graphs and threshold graphs with a square root .
adamaszek and adamaszek @xcite proved that if a graph has a square root of girth at least 6 , then this square root is unique up to isomorphism .
farzad , lau , le and tuy @xcite showed that recognizing graphs with a square root of girth at least @xmath27 is polynomial - time solvable if @xmath28 and np - complete if @xmath29 .
the missing case @xmath30 was shown to be np - complete by farzad and karimi @xcite .
the classical square root problem is a decision problem .
we introduce two optimization variants of it in order to be able to take a _
parameterized _ road to square roots .
a problem with input size @xmath1 and a parameter @xmath0 is said to be _ fixed parameter tractable _ ( or fpt ) if it can be solved in time @xmath31 for some function @xmath32 that only depends on @xmath0 .
we consider two natural choices for the parameter @xmath0 for our optimization variants of the square root problem and in this way obtain the first fpt algorithms for square root problems .
first , in section [ s - min ] , we parameterize the minimum square root problem , which is that of testing whether a graph has a square root with at most @xmath33 edges for some given integer @xmath33 . because any square root of a connected @xmath1-vertex graph @xmath13 is a connected spanning subgraph of @xmath13 , every square root of @xmath13 has at least @xmath34 edges .
consequently , any instance @xmath35 of minimum square root with @xmath36 is a no - instance if @xmath13 is connected , which means that we may assume that @xmath37 .
hence , @xmath38 is the natural choice of parameter .
our main result is that minimum square root is fpt with parameter @xmath0.-vertex graphs with @xmath39 connected components the natural parameter is @xmath40 instead of @xmath38 . because a square root of a graph is the disjoint union of square roots of its connected components , our fpt result for connected graphs immediately carries over to disconnected graphs if we choose as parameter @xmath40 instead . ]
we prove this result by showing that an instance of minimum square root can be reduced to an instance of a more general problem , in which we impose additional requirements on some of the edges , namely to be included or excluded from the square root .
we prove that the new instance has size quadratic in @xmath0 . in other words , we show that minimum square root has a generalized kernel of quadratic size ( see section [ sec : defs ] for the definition of this notion ) .
this result is further motivated by the observation that minimum square root generalizes the problem of recognizing squares of trees ( take @xmath41 ) . a weaker statement of our fpt result is that of saying that the problem of recognizing squares of graphs of the class @xmath42 is polynomial - time solvable for all fixed @xmath5 . as such , our result can also be seen as an extension of the aforementioned result of recognizing squares of trees @xcite .
second , in section [ s - max ] , we parameterize the maximum square root problem , which is that of testing whether a given graph @xmath13 with @xmath2 edges has a square root with at least @xmath33 edges for some given integer @xmath33 .
we show that this problem is fpt with parameter @xmath43 .
this choice of parameter is also natural , as @xmath13 has a square root with at least @xmath33 edges if and only if @xmath13 can be modified into a square root ( of itself ) by at most @xmath0 edge deletions .
hence , our second fpt result can be added to the growing body of parameterized results for graph editing problems , which form a well studied problem area within algorithmic graph theory . in section [ s - max ]
we also present an exact exponential time algorithm for maximum square root , which could be seen as an improvement of the algorithm implied by the characterization of mukhopadhyay @xcite .
in section [ s - con ] we mention a number of relevant open problems .
we only consider finite undirected graphs without loops and multiple edges .
we refer to the textbook by diestel @xcite for any undefined graph terminology and to the textbooks of downey and fellows @xcite , flum and grohe @xcite , and niedermeier @xcite for detailed introductions to parameterized complexity theory .
let @xmath13 be a graph .
we denote the vertex set and edge set of @xmath13 by @xmath9 and @xmath44 , respectively .
the subgraph of @xmath13 induced by a subset @xmath45 is denoted by @xmath46 $ ] .
the graph @xmath47 is the graph obtained from @xmath13 by removing all vertices in @xmath48 . if @xmath49 , we also write @xmath50 .
the _ distance _ @xmath51 between a pair of vertices @xmath11 and @xmath12 of @xmath13 is the number of edges of a shortest path between them .
the _ open neighborhood _ of a vertex @xmath52 is defined as @xmath53 , and its _ closed neighborhood _ is defined as @xmath54 = n_g(u ) \cup \{u\}$ ] .
two vertices @xmath55 are said to be _ true twins _ if @xmath54=n_g[v]$ ] , and @xmath55 are _ false twins _ if @xmath56 . a vertex @xmath11 is _ simplicial _ , if @xmath57 is a clique .
the _ degree _ of a vertex @xmath52 is denoted @xmath58 .
the maximum degree of @xmath13 is denoted @xmath59 .
a vertex of degree 1 is said to be a _ pendant _ vertex .
let @xmath13 be a connected graph .
let @xmath60 , and let @xmath61 and @xmath62 be two disjoint nonempty vertex subsets of @xmath63 .
then @xmath64 is a _ separator _ of @xmath13 if @xmath63 is disconnected , @xmath64 is an _ ( x , y)-separator _ if @xmath63 has no path that connects a vertex of @xmath61 to a vertex of @xmath62 , and @xmath64 is a _ minimal _
@xmath65-separator if @xmath64 is an @xmath65-separator of @xmath13 and no proper subset of @xmath64 is an @xmath65-separator .
moreover , @xmath13 is _ @xmath66-connected _ if and only if @xmath67 and @xmath13 has no separators of size one .
the _ union _ of two graphs @xmath68 and @xmath69 is the graph @xmath70 .
the graph @xmath71 denotes the complete graph on @xmath1 vertices . the graph @xmath72 denotes the star on @xmath73 vertices . a well - known technique to show that a parameterized problem @xmath74 is fixed - parameter tractable is to find a _ reduction to a problem kernel_. this technique replaces an instance @xmath75 of @xmath74 with a reduced instance @xmath76 of @xmath74 called a _ ( problem ) kernel _ such that the following three conditions hold : * @xmath77 and @xmath78 for some computable function @xmath27 ; * the reduction from @xmath75 to @xmath76 is computable in polynomial time ; * @xmath79 is a yes - instance of @xmath74 if and only if @xmath80 is a yes - instance of @xmath74 .
if we slightly modify this definition by letting the instance @xmath80 belong to a different problem than @xmath74 , then @xmath80 is called a _ generalized _ kernel for @xmath74 in the literature .
this concept has been introduced and named _ bikernel _ by alon , gutin , kim , szeider and yeo @xcite ; a related notion is compression .
an upper bound @xmath81 on @xmath82 is called the _ kernel size _ , and a kernel is called _ linear _ if its size is linear in @xmath0 and _ quadratic _ if its size is quadratic in @xmath0 .
it is well known that a parameterized problem is fixed - parameter tractable if and only if it has a kernel ( see for example @xcite ) .
as discussed in section [ s - our ] , we consider connected graphs only and parameterize minimum square root by @xmath38 . from now on we denote this problem as trees @xmath83 edges square root + _ input : _ a connected graph @xmath13 and an integer @xmath5 + _ parameter : _ @xmath0 + _ question : _ has @xmath13 a square root with at most @xmath3 edges ? we show the following result .
[ thm : tree - few - edges ] the tree @xmath83 edges square root problem can be solved in time @xmath84 on graphs with @xmath1 vertices and @xmath2 edges .
the remainder of this section is organized as follows . in section [ s - structural ]
we show a number of structural results needed to prove theorem [ thm : tree - few - edges ] . in section [ s - reduction ]
we consider the more general problem tree @xmath83 edges square root with labels + _ input : _ a connected graph @xmath13 , an integer @xmath5 and two disjoint subsets + @xmath85 + _ parameter : _
+ _ question : _ has @xmath13 a square root @xmath14 with at most @xmath86 edges , such that @xmath87 and @xmath88 ? note that the sets @xmath89 and @xmath90 in this problem are given sets of _ required _ edges ( that have to be in the square root ) and _ blocked _ edges ( that are not allowed to be in the square root ) , respectively . also note that tree @xmath83 edges square root with labels generalizes trees @xmath83 edge square root ; choose @xmath91 . we reduce tree @xmath83 edges square root to tree @xmath83 edges square root with labels where the size of the graph in the obtained instance is @xmath92 .
in other words , we construct a quadratic generalized kernel for tree @xmath83 edges square root .
this means that to solve an instance of trees @xmath83 edge square root , we can solve the obtained instance of tree @xmath83 edges square root with labels by a brute force algorithm . in section [ s - solving ] we analyze the corresponding running time and complete the proof of theorem [ thm : tree - few - edges ] .
we start with the following observation that we will frequently use .
[ obs : leaves ] let @xmath14 be a square root of a connected graph @xmath13 . * if @xmath11 is a pendant vertex of @xmath14 , then @xmath11 is a simplicial vertex of @xmath13 .
* if @xmath55 are pendant vertices of @xmath14 adjacent to the same vertex , then @xmath55 are true twins in @xmath13 . * if @xmath55 are pendant vertices of @xmath14 adjacent to different vertices , then @xmath11 and @xmath12 are not adjacent in @xmath13 unless @xmath93 .
we now state five useful lemmas , the first two of which , lemmas [ lem : leaves - one ] and [ lem : leaves - two ] , can be found implicitly in the paper of ross and harary @xcite .
ross and harary @xcite consider tree square roots , whereas we are concerned with finding general square roots .
as such we give explicit statements of lemmas [ lem : leaves - one ] and [ lem : leaves - two ] .
we also give a proof of lemma [ lem : leaves - two ] ( the proof of lemma [ lem : leaves - one ] is straightforward ) .
[ lem : leaves - one ] let @xmath14 be a square root of a graph @xmath13 .
let @xmath94 for some @xmath95 induce a star in @xmath14 with central vertex @xmath96 .
let @xmath97 be pendant and @xmath98 be a @xmath99-separator of @xmath14 .
then @xmath100 is a clique of @xmath13 , and @xmath101 is a minimal @xmath102-separator of @xmath13 .
[ lem : leaves - two ] let @xmath13 be a connected graph with a square root @xmath14 .
let @xmath100 , @xmath95 be a clique in @xmath13 , such that @xmath101 is a minimal @xmath103-separator of @xmath13 .
let @xmath104 for some @xmath105 and @xmath106 for some @xmath107 , as shown in figure [ fig : trimming ] . then the following three statements hold : * @xmath108 and , either @xmath109 , @xmath110 , @xmath111 @xmath112 , and @xmath98 is a minimal @xmath113-separator in @xmath14 , or @xmath114 , @xmath115 , @xmath116 and @xmath117 is a minimal @xmath118-separator in @xmath14 ( see figure [ fig : trimming - i - ii - iii ] i ) ) .
* if @xmath119 are true twins in @xmath13 , then either @xmath120 or @xmath121 .
moreover , in this case , @xmath13 is the union of two complete graphs with vertex sets @xmath100 and @xmath122 , respectively , and @xmath13 has two ( isomorphic ) square roots with edge sets @xmath123 @xmath124 and @xmath125 @xmath126 , respectively ( see figure [ fig : trimming - i - ii - iii ] ii ) ) . * if @xmath127\setminus n_g[u_1]\neq\emptyset$ ] , then @xmath128 , @xmath110 , @xmath129 .
moreover , the graph @xmath130 obtained from @xmath14 by deleting all @xmath131 with @xmath132 is a square root of @xmath13 ( in which @xmath100 induces a star with central vertex @xmath96 and with leaves @xmath133 that are pendant vertices except for @xmath134 ( see figure [ fig : trimming - i - ii - iii ] iii ) ) .
we first prove i ) .
as @xmath101 is a @xmath103-separator of @xmath13 , at least one vertex @xmath135 with @xmath95 is adjacent to one of @xmath119 in @xmath14 , say to @xmath96 .
then @xmath129 ; otherwise , that is , if @xmath96 is adjacent to some @xmath136 in @xmath14 , then @xmath137 contradicting the fact that @xmath101 is a @xmath103-separator of @xmath13 . because @xmath129
, at least one vertex @xmath138 must be adjacent to @xmath134 in @xmath14 ( as otherwise @xmath14 is not connected and hence can not be the square root of @xmath13 , which is a connected graph ) . because @xmath101 is a @xmath103-separator of @xmath13 , this means that @xmath110 .
consequently , @xmath108 and @xmath98 is a minimal @xmath113-separator in @xmath14 .
suppose that there is a vertex @xmath135 , @xmath139 , such that @xmath140 .
since @xmath97 are not adjacent to @xmath134 , it follows that any @xmath141-path in @xmath14 has length at least 3 , which is not possible as @xmath142 .
we conclude that @xmath143 .
hence we have shown i ) .
we now prove ii ) .
note that @xmath144 with @xmath145 .
due to i ) either @xmath96 or @xmath134 is not adjacent to any @xmath146 . in the first case @xmath134
must be adjacent to all @xmath146 in @xmath14 , as otherwise there is no required path of length at most @xmath66 in @xmath14 between some @xmath146 and @xmath96 .
similarly , in the second case , @xmath96 must be adjacent to all @xmath146 in @xmath14 .
hence , @xmath122 is a clique in @xmath13 .
if @xmath14 has an edge @xmath147 with @xmath148 , then @xmath149 , which is not possible .
this means that @xmath13 is the union of two complete graphs with vertex sets @xmath100 and @xmath122 , respectively .
it is readily seen that @xmath13 has two ( isomorphic ) square roots with edge sets @xmath123 @xmath124 and @xmath125 @xmath126 , respectively .
hence we have shown ii ) .
it remains to prove iii ) .
let @xmath150\setminus n_g[u_1]\neq\emptyset$ ] .
due to i ) we have that @xmath108 , and that either @xmath109 , @xmath110 , @xmath111 @xmath112 , or @xmath114 , @xmath115 , @xmath116 .
if the latter case holds , then any @xmath151-path in @xmath14 has length at least 3 , which is not possible as @xmath152 .
hence the former case must hold .
let @xmath130 be a graph obtained from @xmath14 by deleting all @xmath131 for @xmath153 .
it is readily seen that @xmath154 .
hence we have shown iii ) .
let @xmath13 be a graph that contains ( besides possibly some other vertices ) @xmath155 distinct vertices @xmath156 , @xmath157 , @xmath158 for some @xmath95 , @xmath105 and @xmath107 , such that the following conditions hold : * @xmath100 is a clique in @xmath13 ; * @xmath159 is a minimal @xmath160-separator in @xmath13 if @xmath161 ; * @xmath162 ; * @xmath163 ; * @xmath164 and @xmath165 ; * @xmath166 for @xmath167 and @xmath168 .
we call @xmath13 an _ @xmath169-graph _ and @xmath159 an _
@xmath169-triple _ with _ outer vertices _ @xmath96 and @xmath170 , see figure [ f - fgraph ] for an example .
here , @xmath169 refers to the graph in figure [ fig : path ] .
these notions are further explained by lemmas [ lem : path - one ] and [ lem : path - two ] .
[ lem : path - one ] let @xmath14 be a square root of a graph @xmath13 .
let @xmath14 contain the graph @xmath169 of figure [ fig : path ] as a subgraph , such that @xmath171 are pendant vertices of @xmath14 ( if @xmath161 ) , @xmath172 , @xmath173 is an induced path in @xmath14 that is not contained in any cycle of length at most @xmath174 , @xmath175 and @xmath176 .
then @xmath13 is an @xmath169-graph .
conditions ( i)-(iii ) and ( v ) are readily seen to hold .
conditions iv ) and vi ) follow from the condition that the path @xmath173 is not contained in any cycle of length at most 6 in @xmath14 .
[ lem : path - two ] let @xmath13 be a connected @xmath169-graph . if @xmath14 is a square root of @xmath13 , then the graph @xmath169 of figure [ fig : path ] is a subgraph of @xmath14 such that @xmath172 , @xmath175 and @xmath176 .
moreover , the graph obtained from @xmath14 by deleting all edges @xmath131 with @xmath177 is a square root of @xmath13 that contains @xmath171 as pendant vertices ( if @xmath161 ) .
let @xmath14 be a square root of @xmath13 .
we consider the following three cases . * case 1 .
* @xmath178 .
because @xmath179 , this means that @xmath180 .
symmetrically , @xmath181 .
since each @xmath182 but @xmath183 , @xmath14 has an @xmath184-path of length 2 . because @xmath185 , the middle vertex of this path is in @xmath186 .
because @xmath146 is not adjacent to @xmath97 in @xmath14 ( as it not so in @xmath13 ) , this path goes through @xmath96 . in other words , @xmath187 and , by symmetry , @xmath188 . if a vertex @xmath189 is adjacent to @xmath96 in @xmath14 , then @xmath190 is adjacent to both @xmath134 and @xmath191 in @xmath13 . because @xmath185 , we find that @xmath192 or @xmath193 . however , none of @xmath194 is adjacent to @xmath191 , whereas none of @xmath195 is adjacent to @xmath134 .
we conclude that @xmath175 and by using the same arguments that @xmath176 .
now we show that @xmath196 . to prove it ,
assume that some @xmath135 , @xmath197 , is not adjacent to @xmath134 in @xmath14 .
then @xmath96 and @xmath135 are at distance at least 3 in @xmath14 contradicting @xmath198 .
we already deduced that @xmath180 and that @xmath181 . by assumption , @xmath134 is adjacent to both @xmath96 and @xmath170 . as @xmath185 , we then find that @xmath172 .
to conclude the proof for this case , it remains to observe that if some @xmath199 are adjacent in @xmath14 for @xmath200 , then the graph @xmath130 obtained from @xmath14 by the removal of these edges is a square root of @xmath13 .
* @xmath201 .
since @xmath202 , @xmath203 and @xmath185 , there exists a vertex @xmath204 such that @xmath205 .
because @xmath190 is not adjacent to @xmath206 in @xmath13 , we find that @xmath181 . by the same arguments , we obtain @xmath180 .
hence , @xmath207 . by symmetry ,
some vertex from @xmath208 is adjacent to @xmath170 in @xmath14 .
consequently , each vertex of @xmath159 is adjacent to some vertex in @xmath208 in @xmath14 .
as @xmath159 separates @xmath208 from @xmath209 , this means that @xmath14 has no edges that join @xmath210 with the vertices of @xmath209 ; a contradiction .
hence , this case is not possible . by symmetry , it remains to consider the following case .
* case 3 . * @xmath108 and @xmath211 .
because @xmath108 and @xmath212 , we find that @xmath181 . because @xmath213 , this means that @xmath14 contains a @xmath214-path of length 2 .
because @xmath211 and @xmath185 , such a path should go through one of the vertices of @xmath215 .
however , none of these vertices is adjacent to @xmath216 in @xmath13 , and consequently not in @xmath14 either ; a contradiction .
therefore , this case is not possible either .
[ lem : twins ] let @xmath55 be true twins in a connected graph @xmath13 with at least three vertices .
let @xmath217 be the graph obtained from @xmath13 by deleting @xmath12 .
the following two statements hold : * if @xmath130 is a square root of @xmath217 , then the graph @xmath14 obtained from @xmath130 by adding @xmath12 with @xmath218 ( that is , by adding a false twin of @xmath11 ) is a square root of @xmath13 . * if @xmath14 is a square root of @xmath13 such that @xmath55 are false twins in @xmath14 , then the graph @xmath130 obtained by deleting @xmath12 is a square root of @xmath217 .
we first prove i ) .
let @xmath130 be a square root of @xmath217 , and let @xmath14 be the graph obtained from @xmath130 by adding a false twin @xmath12 of @xmath11 . as @xmath13 is a connected graph with at least three vertices , @xmath11 is adjacent in @xmath130 to some vertex @xmath190
. then @xmath11 and @xmath12 are adjacent to @xmath190 in @xmath14 and thus @xmath219 .
hence , @xmath220 is an edge of @xmath221 .
then it is straightforward to see that @xmath15 .
statement ii ) follows from the fact that identifying false twins does not change the distance between any two vertices .
as discussed , in this section , we reduce tree @xmath83 edges square root to tree @xmath83 edges square root with labels in such a way that the size of the graph in the obtained instance is @xmath92 .
first , we informally sketch the main steps of the reduction .
let @xmath13 be a connected graph with @xmath1 vertices , and let @xmath0 be a positive integer .
suppose that @xmath14 is a square root of @xmath13 with at most @xmath222 edges .
if @xmath14 has a vertex @xmath11 of degree at least 2 that has exactly one non - pendant neighbor @xmath12 , then we recognize the corresponding structure in @xmath13 and delete those vertices of @xmath13 that are pendant vertices of @xmath14 adjacent to @xmath11 as shown in figure [ fig : trim - informal ] , that is , similar to the algorithm of lin and skiena @xcite , we `` trim '' pendant edges in potential roots . since the root we are looking for is not a tree ,
our trimming is more sophisticated and based on lemmas [ lem : leaves - one ] and [ lem : leaves - two ] .
we will show that in this way we obtain a graph @xmath217 with @xmath223 vertices that has the following property : every pendant vertex of any square root @xmath130 of @xmath217 with at most @xmath224 edges is adjacent to a vertex that has at least two non - pendant neighbors in @xmath130 .
suppose that @xmath130 has a sufficiently long induced path @xmath225 , such that every internal vertex of @xmath225 has exactly two non - pendant neighbors in @xmath130 .
let @xmath11 be an internal vertex of @xmath225 , and let @xmath226 be the two non - pendant neighbors of @xmath11 .
using lemmas [ lem : path - one ] and [ lem : path - two ] , we recognize the corresponding structure in @xmath217 and modify @xmath217 as shown in figure [ fig : path - informal ] , that is , we delete @xmath11 in @xmath130 and join @xmath227 any @xmath228 by an edge . by performing this operation recursively , we obtain a graph @xmath229 with @xmath230 vertices .
suppose that @xmath231 is a square root of @xmath229 with at most @xmath232 edges .
let @xmath233 be the graph obtained from @xmath231 by deleting all pendant vertices of @xmath231 .
then @xmath233 has no vertices of degree 1 , and the length of every path @xmath225 with internal vertices of degree 2 in @xmath233 is bounded by a constant .
this means that the size of @xmath233 is @xmath234 .
the vertices of @xmath235 are pendant vertices of @xmath231 .
consider the set @xmath236 of pendant vertices of @xmath231 adjacent to a vertex @xmath237 .
then the vertices of @xmath236 are simplicial vertices of @xmath229 . moreover
, they are true twins .
we use observation [ obs : leaves ] and lemma [ lem : twins ] to show that we may reduce the number of true twins in @xmath229 if @xmath229 has too many .
this results in a graph @xmath238 with @xmath239 vertices such that @xmath239 is @xmath92 . during the reduction from @xmath13 to @xmath229
we label some edges , that is , we include some edges in sets @xmath89 or @xmath90 and , therefore , obtain an instance @xmath240 of tree @xmath83 edges square root with labels .
before we give a formal description of our reduction , we introduce the following terminology .
a square root @xmath14 of a graph @xmath13 that has at most @xmath241 edges for some @xmath5 is called a _ solution _ of the instance @xmath242 of tree @xmath83 edges square root . if @xmath87 and @xmath88 for two disjoint subsets @xmath89 and @xmath90 of @xmath44 , then @xmath14 is also called a _
solution _ of the instance @xmath243 of tree @xmath83 edges square root with labels .
we are now ready to give the exact details of our reduction .
let @xmath13 be a connected graph with @xmath1 vertices and @xmath2 edges , and let @xmath0 be a positive integer .
first we check whether @xmath13 has a square root that is a tree by using the linear - time algorithm of lin and skiena @xcite .
if we find such a square root , then we stop and return a yes - answer .
from now we assume that every square root of @xmath13 ( if there exists one ) has at least one cycle .
because connected graphs that have square roots are 2-connected , we also check whether @xmath13 is 2-connected .
if so , then we stop and return a no - answer .
otherwise we continue as follows .
we introduce two sets of edges @xmath89 and @xmath90 .
initially , we set @xmath91 .
next , we `` trim '' pendant edges in potential roots , that is , we exhaustively apply the following rule that consists of five steps that must be performed in increasing order . * trimming rule * 1 .
find a pair @xmath244 of two adjacent vertices such that one connected component of @xmath63 consists of @xmath95 vertices @xmath97 that together with @xmath119 form a clique in @xmath13 .
if @xmath245=n_g[u_2]$ ] then stop and return a no - answer .
3 . if @xmath245\setminus n_g[u_2]\neq\emptyset$ ] and @xmath127\setminus n_g[u_1]\neq\emptyset$ ] , then stop and return a no - answer .
4 . if @xmath245\setminus n_g[u_2]\neq\emptyset$ ] , then rename @xmath96 by @xmath134 and @xmath134 by @xmath96 ( this step is for notational convenience only and has no further meaning ) .
define sets @xmath246 and @xmath247 .
if @xmath248 or @xmath249 , then stop and return a no - answer .
otherwise , set @xmath250 , @xmath251 , delete @xmath97 from @xmath13 and also delete all edges incident to @xmath97 from @xmath89 and @xmath90 . exhaustively applying the trimming rule yields a sequence of instances @xmath252 of tree @xmath83 edges square root with labels for some integer @xmath253 , where @xmath254 and where @xmath255 is an instance for which we have either returned a no - answer ( in steps 2 , 3 or 6 ) or for which there does not exist a set @xmath64 as specified in step 1 .
for @xmath256 we denote the sets @xmath257 and @xmath258 constructed in the @xmath259th call of the trimming rule by @xmath260 and @xmath261 , respectively .
we need the following lemma .
[ lem : trimming ] the instance @xmath262 has no solution that is a tree , and @xmath263 is @xmath66-connected .
moreover , @xmath255 has a solution if and only if @xmath264 has a solution .
if the trimming rule returned a no - answer for @xmath255 , then @xmath264 has no solution . for @xmath265
, we use induction to show that the graph @xmath266 is @xmath66-connected and that @xmath267 has no solution that is a tree .
moreover , for all @xmath268 , we show that @xmath269 has a solution if and only if @xmath270 has a solution .
finally , we prove that if the trimming rule returned a no - answer for @xmath255 , then @xmath264 has no solution . if @xmath271 , then @xmath266 is 2-connected and @xmath272 has no solution that is a tree by our initial assumption ( as we had preprocessed @xmath13 with respect to these two properties ) .
now suppose that @xmath268 . by our induction hypothesis
, we may assume that @xmath273 is 2-connected and that @xmath274 has no solution that is a tree .
because the trimming rule applied on @xmath270 yielded a new instance @xmath269 , the graph @xmath273 has a pair @xmath244 of adjacent vertices such that one connected component of @xmath63 consists of vertices @xmath97 that together with @xmath119 form a clique in @xmath273 .
step 6 implies that @xmath275 .
because we did not return a no - answer for @xmath270 , we find that @xmath276 \subset n_{g_{i-1}}[u_2]$ ] .
hence , @xmath273 is not a complete graph .
because @xmath273 is 2-connected , this means that @xmath266 is 2-connected .
we now show that any solution for @xmath274 corresponds to a solution for @xmath267 , and vice versa .
first suppose that @xmath277 is an arbitrary solution for @xmath274 .
let @xmath278 .
because @xmath279\setminus n_{g_{i-1}}[u_1]\neq\emptyset$ ] , we find that @xmath280 contains at least two connected components .
as @xmath273 is 2-connected , this means that @xmath101 is a minimal @xmath281-separator of @xmath273 .
hence we may apply lemma [ lem : leaves - two ] iii ) , which tells us that @xmath282 , @xmath283 , and @xmath284 . as @xmath285 and @xmath286 , this means that the graph obtained from @xmath277 by deleting @xmath97 is a solution for @xmath269 ; in particular note that @xmath287 , as required .
now suppose that @xmath288 is an arbitrary solution for @xmath269 . then adding the edges @xmath289 to @xmath288 yields a graph @xmath14 that is a square root of @xmath273 . the edges @xmath289 are not in @xmath290 , as they are in the set
@xmath291 constructed in step 5 and @xmath292 ( otherwise the trimming rule would have stopped when processing @xmath270 in step 6 ) .
now suppose that @xmath293 contains an edge not in @xmath14 . by definition of @xmath294
, this edge must be between some @xmath295 and @xmath296 with @xmath297 .
then @xmath298 belongs to @xmath294 , because it was placed in the set @xmath299 for some @xmath300 . in step 4 of the corresponding call of the trimming rule , also one of the edges @xmath301 or @xmath302
was placed in @xmath303 .
hence either @xmath301 or @xmath302 belongs to @xmath290 .
this yields a contraction as both @xmath301 and @xmath302 belong to @xmath291 and @xmath292 ( otherwise the trimming rule would have stopped when processing @xmath270 in step 6 ) .
hence , after observing that @xmath304 , we conclude that @xmath14 is a solution for @xmath270 .
we observe that @xmath288 can not be a tree , as this would imply that @xmath14 is a tree , which is not possible as @xmath270 does not have such a solution .
we are left to show that if the trimming rule returned a no - answer for @xmath255 , then @xmath264 has no solution . due to the above
, this comes down to showing that @xmath255 has no solution .
suppose that the trimming rule returned a no - answer for @xmath255 .
then this must have happened in step 2 , 3 or 6 , thus after step 1 .
hence , there exists a pair of adjacent vertices @xmath244 in @xmath263 , such that one connected component of @xmath305 has vertex set @xmath194 and @xmath100 is a clique .
first assume that @xmath64 is not a separator of @xmath263 , that is , @xmath263 is a complete graph with vertex set @xmath100
. then @xmath245=n_g[u_2]$ ] ( and the no - answer given by the trimming rule happens in step 2 ) .
in order to obtain a contradiction , assume that @xmath255 has a solution @xmath14 .
any star on @xmath306 vertices is a square root of @xmath263 with at most @xmath307 edges .
however , @xmath14 can not be such a star , as @xmath255 has no solution that is a tree .
hence , @xmath308 or @xmath309 .
recall that @xmath310 .
hence , @xmath311 , and non - emptiness of @xmath312 or @xmath313 must have been obtained in a previous call of the trimming rule , say in the @xmath314th call of the trimming rule for some @xmath315 . by definition of steps 5 and 6 , we find that @xmath316 implies that @xmath317 .
hence , @xmath317 . let @xmath318 . by steps 5 and 6 , this edge has an end - vertex , say @xmath135 , such that @xmath319 for all @xmath320 . consequently , @xmath321 for all @xmath322 .
because the star with central vertex @xmath323 and leaves @xmath324 is not a solution for @xmath255 , there must be an edge @xmath325 with @xmath326 .
however then , due to steps 5 and 6 , @xmath327 or @xmath328 , that is , at least one of these edges can not be in @xmath14 ; a contradiction .
now assume that @xmath64 is a separator of @xmath263 .
because @xmath263 is 2-connected , both @xmath96 and @xmath134 have at least one neighbor in @xmath329 .
hence @xmath101 is a minimal separator ( and we may apply lemma [ lem : leaves - two ] in the remainder ) . recall that the trimming rule only returns a no - answer in steps 2 , 3 , or 6 .
we consider each of these three cases separately .
* case 1 . *
the no - answer is given in step 2 .
then @xmath245=n_g[u_2]$ ] . by lemma [ lem :
leaves - two ] i ) and ii ) , @xmath263 is the union of two cliques @xmath100 and @xmath330 where @xmath331 . in order to obtain a contradiction ,
suppose that @xmath255 has a solution @xmath14 .
by lemma [ lem : leaves - two ] i ) and ii ) , we may assume without loss of generality that @xmath332 , @xmath333 , @xmath129 and @xmath121 .
recall that @xmath255 has no solution that is a tree .
hence , there exists an edge @xmath334 for some @xmath335 or an edge @xmath336 for some @xmath337 . by symmetry
, we only need to consider the case @xmath334 .
this edge was placed in @xmath312 in some previous call of the trimming rule . however , due to steps 5 and 6 performed in that call , we find that @xmath338 or @xmath339 , that is , at least one of these two edges can not be in @xmath14 ; a contradiction .
* case 2 . *
the no - answer is given in step 3 .
then we have @xmath245\setminus n_g[u_2]\neq\emptyset$ ] and @xmath127\setminus n_g[u_1]\neq\emptyset$ ] .
due to lemma [ lem : leaves - two ] i ) and iii ) , @xmath255 has no solution .
* case 3 . *
the no - answer is given in step 6 .
then @xmath340 or @xmath341 .
by step 4 , we may assume that @xmath245\setminus n_g[u_2]=\emptyset$ ] and that @xmath127\setminus n_g[u_1]\neq\emptyset$ ] . in order to obtain a contradiction ,
suppose that @xmath255 has a solution @xmath14 .
by lemma [ lem : leaves - two ] iii ) , @xmath342 .
hence @xmath343 , which means that @xmath340 .
let @xmath104 .
then we have that @xmath344 . by the same arguments as used in case 1 , we find that @xmath345 for all @xmath346 . by lemma [ lem :
leaves - two ] iii ) , we find that @xmath347 , and hence @xmath312 , does not contain the edges @xmath348 .
we conclude that @xmath349 ; a contradiction .
lemma [ lem : trimming ] shows that the trimming rule is safe , that is , we either found that @xmath350 has no solution , or that we may continue with the instance @xmath255 instead .
suppose the latter case holds .
recall that @xmath255 has no set @xmath64 as specified in step 1 , as otherwise we would have applied the trimming rule once more . to simplify notation
, we write @xmath351 .
we need the following properties that hold for every solution of @xmath243 ( should @xmath243 have a solution ) .
[ lem : trimmed ] any solution @xmath14 of @xmath243 satisfies the following properties : * the neighbor of every pendant vertex of @xmath14 has at least two non - pendant neighbors in @xmath14 ; * only edges of @xmath13 incident to pendant vertices of @xmath14 can be in @xmath89 or @xmath90 ; * if a pendant vertex @xmath12 of @xmath14 is incident to an edge of @xmath89 in @xmath13 , then all other edges of @xmath13 that are incident to @xmath12 are in @xmath90 . in order to show i ) , suppose that @xmath14 is a solution of an instance @xmath243 , such that @xmath14 contains a pendant vertex @xmath11 adjacent to a vertex @xmath12 . if @xmath352 , then @xmath14 is isomorphic to @xmath353 , which is not possible as @xmath243 has no solution that is a tree .
hence @xmath354 and @xmath12 has at least one neighbor other than @xmath11 .
if all neighbors of @xmath12 are pendant , then @xmath14 is a tree ; a contradiction .
hence , @xmath12 has at least one non - pendant neighbor .
if @xmath12 has a unique non - pendant neighbor @xmath355 , then by lemma [ lem : leaves - one ] , @xmath356 contains a connected component induced by the pendant neighbors of @xmath12 whose vertices together with @xmath12 and @xmath355 form a clique in @xmath13 .
hence , we can apply the trimming rule on @xmath357 , which is a contradiction .
properties ii ) and iii ) follow from the construction of @xmath89 and @xmath90 in steps 4 and 5 of the trimming rule .
we now exhaustively apply the following rule on @xmath243 .
this rule consists of four steps that must be performed in increasing order .
* path reduction rule * 1 .
find an @xmath169-triple @xmath358 .
2 . set @xmath359 and @xmath360 ( note that the set @xmath361 if @xmath362 ) .
3 . if @xmath248 or @xmath249 , then stop and return a no - answer .
4 . delete @xmath363 from @xmath13 .
delete all edges incident to @xmath363 from @xmath89 and @xmath90 .
if @xmath364 , then delete @xmath365 from @xmath90 .
add @xmath365 to @xmath89 .
add @xmath366 and @xmath367 in @xmath13 .
put these edges in @xmath90 . exhaustively applying the path reduction rule yields a sequence of instances @xmath252 of tree @xmath83 edges square root with labels for some integer @xmath253 , where @xmath368 and where @xmath255 is an instance for which we have either returned a no - answer ( in step 3 ) or for which there does not exist an @xmath169-triple @xmath64 .
for @xmath265 we denote the sets @xmath257 and @xmath258 constructed in the @xmath259th call of the path reduction rule by @xmath260 and @xmath261 , respectively .
we need the following lemma , which we will use at several places .
[ l - back ] let @xmath268 and @xmath159 be the @xmath169-triple that yielded instance @xmath269 . if @xmath288 is a solution for @xmath269 , then @xmath369 and the graph @xmath277 obtained from @xmath288 by removing the edge @xmath365 and by adding @xmath134 and vertices @xmath171 ( if @xmath161 ) together with edges @xmath370 is a solution for @xmath371 .
we find that @xmath365 is an edge in @xmath288 , because @xmath372 due to step 4 of the last call of the path reduction rule .
the graph @xmath277 is not only a square root of @xmath273 but even a solution for @xmath270 for the following reasons .
first , @xmath277 has at most @xmath373 edges .
second , @xmath277 contains no edge of @xmath290 as the added edges @xmath370 are all in @xmath374 and @xmath375 .
third , @xmath277 contains all the edges of @xmath293 , which can be seen as follows .
suppose that @xmath277 misses an edge of @xmath293
. then this edge must be in @xmath376 . however , this set is equal to @xmath377 and @xmath378 .
we also need the following lemma about true twins in @xmath379 that we will use later as well .
[ l - twins ] let @xmath268 and @xmath159 be the @xmath169-triple that yielded instance @xmath269 . then
any true twins @xmath380 in @xmath266 are true twins in @xmath273 .
suppose that @xmath266 has true twins @xmath380 that are not true twins in @xmath273 .
consider the corresponding @xmath169-graph that yielded the instance @xmath269 . because @xmath381 are not true twins in @xmath273 , the neighborhood of @xmath12 or @xmath355 is modified by the path reduction rule .
we may assume without loss of generality that the neighborhood of @xmath12 is changed .
note that neither @xmath382 nor @xmath383 if @xmath95 , because these vertices have been removed in step 4 of the path reduction rule when @xmath266 was constructed . as @xmath384 either
, we find that @xmath385 . by symmetry
we may assume that @xmath386 .
we observe that @xmath12 is adjacent to both @xmath96 and @xmath170 in @xmath266 .
because the neighborhood of each @xmath146 is modified in the same way ( namely by the removal of @xmath134 and the addition of @xmath170 ) , we find that @xmath387 . because @xmath11 and @xmath12 are true twins , they are adjacent . because no two vertices @xmath146 and @xmath388 are adjacent in @xmath266
, we then obtain that @xmath389 .
we conclude that the neighborhood of @xmath355 is not modified by the application of the path reduction rule .
because @xmath12 is adjacent to @xmath96 and @xmath170 in @xmath266 and @xmath381 are true twins in @xmath266 , this means that @xmath355 is adjacent to @xmath96 and @xmath170 in @xmath273 already .
however , by definition of an @xmath169-graph , @xmath390 , and @xmath363 are not in @xmath266 as they were removed by the path reduction rule ; a contradiction .
the next lemma is the analog of lemma [ lem : trimming ] for the path reduction rule .
[ l - reduction ] the instance @xmath262 has no solution that is a tree , and @xmath263 is @xmath66-connected .
moreover , @xmath255 has a solution if and only if @xmath264 has a solution .
if the path reduction rule returned a no - answer for @xmath255 , then @xmath264 has no solution . for @xmath265
, we use induction to show that the graph @xmath266 is @xmath66-connected and that @xmath267 has no solution that is a tree .
moreover , for all @xmath268 , we show that @xmath269 has a solution if and only if @xmath270 has a solution . finally , we prove that if the path reduction rule returned a no - answer for @xmath255 , then @xmath264 has no solution .
if @xmath271 , then @xmath266 is 2-connected and @xmath272 has no solution that is a tree by lemma [ lem : trimming ] .
now suppose that @xmath268 . by our induction hypothesis
, we may assume that @xmath273 is 2-connected and that @xmath274 has no solution that is a tree . because the path reduction rule applied on @xmath270 yielded a new instance @xmath269 , the graph @xmath273 has an @xmath169-triple @xmath358 .
because @xmath273 is 2-connected , @xmath266 is 2-connected ; in particular note that @xmath105 and @xmath107 by definition of an @xmath169-triple .
first suppose that @xmath277 is a solution for @xmath270 .
we claim that @xmath277 contains no edge @xmath391 with @xmath392 .
we prove this claim by contradiction : let @xmath393 for some @xmath392 .
suppose that @xmath394 .
we may apply lemma [ lem : trimmed ] as @xmath264 has a solution @xmath395 ; if @xmath396 this fact follows from the induction hypothesis . by lemma [ lem : trimmed ]
we find that either @xmath295 is a pendant vertex in @xmath395 with @xmath296 as its ( unique ) neighbor , or the other way around .
we may assume without loss of generality that the first case holds , that is , @xmath295 is pendant in @xmath395 and has @xmath296 as its neighbor . note that @xmath397\subseteq n_{g_0}[u_t]$ ] .
we claim that @xmath398\subseteq n_{g_h}[u_t]$ ] for all @xmath399 . to obtain a contradiction ,
suppose not .
then at some point @xmath295 will be made adjacent to a vertex @xmath12 not adjacent to @xmath296 for the first time in step 4 of some call of the path reduction rule .
let @xmath400 be the corresponding @xmath169-triple .
then we may assume without loss of generality that either @xmath401 is adjacent to @xmath402 and @xmath403 but not to @xmath404 , or that @xmath405 is adjacent to @xmath406 but not to @xmath407 .
in the first case , @xmath296 is not in @xmath408 , but must be adjacent to @xmath402 and @xmath403 by our assumption , and hence , the edge @xmath409 will be added in the same step ; a contradiction . in the second case ,
as @xmath295 is adjacent to @xmath402 and @xmath403 , also @xmath296 is adjacent to @xmath402 and @xmath403 ( again by our assumption ) . because @xmath296 does not get removed in this step ( as @xmath296 belongs to @xmath273 ) , this violates the definition of an @xmath169-triple .
we conclude that @xmath398\subseteq n_{g_h}[u_t]$ ] for all @xmath399 .
we first assume that @xmath410 is an edge in @xmath411 .
step 4 of the path reduction rule only moves an edge @xmath412 from a @xmath90-set to an @xmath89-set if @xmath402 and @xmath413 are outer vertices of an @xmath169-triple . in that case
all their common neighbors will be removed from the graph by the definition of an @xmath169-triple . because @xmath398\subseteq n_{g_h}[u_t]$ ] for all @xmath399 , we find that @xmath296 is a common neighbor of @xmath134 and @xmath295 in @xmath414 for all @xmath399 ; in particular @xmath296 belongs to @xmath273 .
hence , the edge @xmath410 will never be moved from @xmath303 to @xmath299 in step 4 of the @xmath314th call of the path reduction rule for some @xmath399 .
if @xmath410 is not an edge in @xmath411 , then at some point it will be an edge due to step 4 of some call of the path reduction rule , say the @xmath415th call for some @xmath416 . in the same step , @xmath410 will be placed in the set @xmath417 .
then , again because @xmath398\subseteq n_{g_h}[u_t]$ ] for all @xmath399 , the edge @xmath410 will never be moved from @xmath417 to a set @xmath299 for some @xmath418 .
hence , in both cases , we find that @xmath419 even if @xmath396 .
as @xmath420 ( due to step 2 in the @xmath421th call ) , we find that @xmath422 .
hence , the path reduction rule would return a no - answer for @xmath270 in step 3 , and consequently the instance @xmath269 would not exist ; a contradiction . now suppose that @xmath298 was placed in some set @xmath299 for some @xmath423 .
properties ii ) and iii ) of an @xmath169-graph together with step 4 of the path reduction rule imply the following : if @xmath295 and @xmath296 form a triangle with some vertex @xmath190 , then @xmath424 or @xmath425 . moreover , in the case in which @xmath426 , this property is not violated by any subsequent intermediate calls of the path reduction rule .
hence , if @xmath391 , then @xmath419 or @xmath427 , and as @xmath428 as well , we derive the same contradiction as before . we conclude that @xmath277 contains no edge @xmath391 with @xmath392 .
also , by lemma [ lem : path - two ] , we may assume without loss of generality that @xmath277 contains no edge @xmath429 with @xmath392 ; otherwise we could remove such an edge from @xmath277 , and the resulting graph would still be a solution for @xmath270 .
consequently , @xmath171 are pendant vertices of @xmath277 .
this means that the graph @xmath14 obtained from @xmath277 by deleting vertices @xmath363 and adding the edge @xmath365 is not only a square root of @xmath266 with at most @xmath430 edges but even a solution for @xmath269 .
now suppose that @xmath288 is a solution for @xmath269 . by lemma [ l - back ] , the graph @xmath14 obtained from @xmath288 by removing the edge @xmath365 and by adding @xmath134 and vertices @xmath171 ( if @xmath161 ) together with edges @xmath370 is a solution for @xmath371 .
we observe that @xmath288 can not be a tree , as this would imply that @xmath14 is a tree , which is not possible as @xmath270 does not have such a solution by the induction hypothesis .
finally , suppose that the path reduction rule returned a no - answer for @xmath255 .
we must show that @xmath264 has no solution . due to the above this comes down to showing that @xmath255 has no solution .
the only step in which the path reduction rule can return a no - answer is in step 3 , meaning that @xmath263 has an @xmath169-triple @xmath358 such that @xmath431 or @xmath432 . in order to obtain a contradiction ,
suppose that @xmath255 has a solution @xmath14 .
by lemma [ lem : path - two ] , the graph @xmath169 shown in figure [ fig : path ] is a subgraph of @xmath14 such that @xmath172 , @xmath175 and @xmath176 .
consequently , @xmath433 , and hence @xmath343 , and moreover , @xmath434 , and hence @xmath435 ; a contradiction .
lemma [ l - reduction ] shows that the path reduction rule is safe , that is , we either found that @xmath264 has no solution , or that we may continue with the instance @xmath255 instead .
suppose the latter case holds .
recall that @xmath255 has no @xmath169-triple , as otherwise we would have applied the path reduction rule once more . also recall that @xmath436 is the set of vertices in the set @xmath89 immediately after the trimming rule .
we write @xmath437 and @xmath438 . to simplify notation , from now on ,
we also write @xmath351 ; note that @xmath439 .
we need the following properties that hold for every solution of @xmath243 ( should @xmath243 have a solution ) .
we call an induced cycle @xmath440 in a graph @xmath14 _ semi - pendant _ if all but at most one of the vertices of @xmath440 are only adjacent to pendant vertices of @xmath14 and their neighbors on @xmath440 . similarly , we call an induced path @xmath225 in a graph @xmath14 _ semi - pendant _ if all internal vertices of @xmath225 are only adjacent to pendant vertices of @xmath14 and their neighbors on @xmath225 .
[ lem : reduced ] any solution @xmath14 of @xmath243 has the following properties : * the neighbor of every pendant vertex of @xmath14 has at least two non - pendant neighbors in @xmath14 ; * only edges of @xmath13 incident to pendant vertices of @xmath14 can be in @xmath441 , and if a pendant vertex @xmath12 of @xmath14 is incident to an edge of @xmath89 , then all other edges of @xmath13 that are incident to @xmath12 are in @xmath90 ; * no edge of @xmath442 is incident to a pendant vertex of @xmath14 ; * the length of every semi - pendant path in @xmath14 is at most @xmath443 ; * the length of every semi - pendant cycle in @xmath14 is at most @xmath174 .
we prove that property i ) holds by contradiction .
suppose that @xmath14 contains a vertex @xmath12 that is the ( unique ) neighbor of a pendant vertex @xmath11 , such that @xmath12 has at most one non - pendant neighbor in @xmath14 .
if all neighbors of @xmath12 in @xmath14 are pendant , then @xmath14 is a tree .
however , this would contradict lemma [ l - reduction ] .
hence , @xmath12 has a unique non - pendant neighbor in @xmath14 .
recall that @xmath14 is a solution for @xmath255 .
note that if @xmath12 is an outer vertex of the corresponding @xmath169-triple , then lemma [ l - back ] tells us that @xmath444 has a solution @xmath445 in which @xmath12 is a non - pendant vertex that has at least one pendant neighbor and that has a unique non - pendant neighbor .
hence , by applying lemma [ l - back ] inductively , we obtain that @xmath264 has a solution @xmath395 containing a vertex with exactly the same property .
this contradicts lemma [ lem : trimmed ] i ) .
we conclude that property i ) holds .
we now show property ii ) . by lemma [ lem : trimmed ] , every edge of @xmath411 that is in @xmath436 is incident to a pendant vertex @xmath11 of any solution for @xmath264 such that all the other edges of @xmath11 belong to @xmath446 .
we observe that , when applying the path reduction rule , @xmath11 will neither be in an @xmath169-triple nor removed from the graph , but @xmath11 could be a vertex of @xmath227-type or @xmath228-type . hence , the path reduction rule may change the neighbors of @xmath11 but if so any new edges incident to it will be placed in @xmath90 ( and stay in @xmath90 afterward ) .
consequently , @xmath11 must be a pendant vertex in any solution for @xmath447 ) as well .
we conclude that ii ) holds .
we now prove prove property iii ) .
recall that we applied the path reduction rule only after first applying the trimming rule exhaustively .
when we apply the path reduction rule on an @xmath169-triple @xmath159 , then afterward @xmath96 and @xmath170 have degree at least 2 in any solution for the resulting instance , which can be seen as follows .
the edge @xmath365 is added to @xmath448 , and hence belongs to any solution .
we also have that @xmath96 is adjacent to @xmath191 in @xmath13 , whereas the edge @xmath449 belongs to @xmath90 .
this means that @xmath96 can not be made adjacent to @xmath191 via the path @xmath450 in @xmath14 , and as such must have at least one other neighbor in @xmath14 . for the same reason @xmath170 , which is adjacent to @xmath216 in @xmath13 whereas @xmath451 , must have another neighbor in @xmath14 besides @xmath96 . as a consequence ,
any edge in @xmath442 can not be incident to a pendant vertex of @xmath14 , that is , we have shown property iii ) .
we now prove property iv ) .
let @xmath225 be a semi - pendant path of length at least 6 in @xmath14 . by definition
, @xmath225 is an induced path .
hence , we can take any three consecutive vertices of @xmath225 as the three vertices @xmath210 in lemma [ lem : path - one ] . by applying this lemma , we find that @xmath13 is an @xmath169-graph implying that we could have applied the path reduction rule once more ; a contradiction .
property v ) can be proven by using the same arguments .
we need the following lemma that holds in case a solution exists for @xmath243 .
[ l - hstarbounded ] the number of non - pendant vertices of any solution for @xmath243 is at most @xmath452 .
suppose @xmath243 has a solution @xmath14 .
let @xmath236 be the set of pendant vertices of @xmath14 , and let @xmath453 .
we need to show that @xmath454 has at most @xmath452 vertices .
let @xmath455 be the set of vertices that have degree at least @xmath456 in @xmath233 , and let @xmath457 be the set of vertices of degree 2 in @xmath233 . by lemma [ lem : reduced ] i )
every vertex of @xmath14 that is adjacent to a pendant vertex of @xmath14 has degree at least @xmath66 in @xmath233 .
hence , @xmath233 has no vertices of degree at most 1 , that is , @xmath458 . because @xmath14 is a solution for @xmath243 , we have that @xmath459 .
this means that @xmath460 & = & |e_{h^*}|\\[1 mm ] & = & \frac{1}{2}\sum_{v}d_{h^*}(v)\\[2 mm ] & \geq & \frac{1}{2}(3|v'|+2|v''| ) .
\end{array}\ ] ] hence , @xmath461 . let @xmath462 be the number of paths in @xmath233 that only have internal vertices of degree 2 ; note that by lemma [ lem : reduced ] iv ) the length of such paths is at most 5 .
let @xmath463 be the number of cycles in @xmath233 that have exactly one vertex of degree at least 3 ; note that by lemma [ lem : reduced ] v ) the length of such cycles is at most 6 . because @xmath464 , we find that @xmath465 and that @xmath466
. hence , @xmath467 .
consequently , @xmath233 has at most @xmath468 vertices .
we are now ready to state our final reduction rule .
the goal of this rule is to apply it once in order to deduce either that @xmath243 has no solution or to derive a new instance of bounded size .
a _ true twin partition _ of a set of vertices @xmath64 of a graph @xmath13 is a partition @xmath469 of @xmath64 such that for all @xmath470 and all @xmath471 we have that @xmath11 and @xmath12 are in @xmath472 if and only if @xmath11 and @xmath12 are true twins in @xmath13 .
if @xmath64 consists of simplicial vertices only we observe that there is no edge between any two vertices that belong to different sets @xmath472 and @xmath473 .
* simplicial vertex reduction rule * 1 .
find the set @xmath64 of all simplicial vertices of @xmath13 that are not incident to the edges of @xmath442 , and moreover , that have all but one of their incident edges in @xmath90 should they be incident to an edge of @xmath441 .
2 . if @xmath474 , then stop and return a no - answer .
3 . construct the true twin partition @xmath469 of @xmath64 .
let @xmath475 be the sets of vertices incident to an edge of @xmath441 in @xmath469 , respectively .
4 . if @xmath476 , then stop and return a no - answer .
5 . if there exist a set @xmath477 such that the edges of @xmath441 incident to a vertex of @xmath477 have no common end - vertex , then stop and return a no - answer .
if there exist a set @xmath472 such that @xmath478 and such that there are three vertices @xmath479 , @xmath480 and @xmath481 with @xmath482 and @xmath483 , then stop and return a no - answer .
7 . for @xmath484 , if @xmath485 , then take @xmath486 arbitrary vertices of @xmath477 and delete them both from @xmath13 and from @xmath472 , also delete the edges of @xmath89 and @xmath90 that are incident to these vertices .
8 . for @xmath484 ,
if @xmath487 , then delete @xmath488 arbitrary vertices of @xmath489 from @xmath13 , also delete the edges of @xmath89 and @xmath90 that are incident to these vertices . applying the simplicial vertex reduction rule on @xmath243 either yields a no - answer ( in step 2 , 4 , 5 or 6 ) or a new instance @xmath490 of tree @xmath83 edges square root with labels .
we will show that if @xmath491 exists , then its size is bounded by a quadratic function of @xmath0 .
for doing so we first need the following two lemmas .
[ l - claim1 ] for @xmath484 , no vertex of @xmath489 is incident to an edge in @xmath89 . by definition of @xmath472 , no vertex of @xmath472 , and hence no vertex of @xmath489 , is incident to an edge in @xmath442 . by definition of @xmath477 , no vertex in @xmath489 is incident to an edge in @xmath441 . because @xmath439 , we have proven lemma [ l - claim1 ] . for @xmath492 , we let @xmath493 denote the set of edges of @xmath90 incident to @xmath227 .
[ l - claim2 ] @xmath494 for all @xmath495 .
let @xmath495 and let @xmath496 for some @xmath497 .
we first show that @xmath498 and we then prove that @xmath499 . in order to obtain a contradiction ,
assume that @xmath500 .
then @xmath501 was included in @xmath90 either by an application of the trimming rule or by an application of the path reduction rule . in both cases ,
@xmath501 was also made adjacent to an edge of @xmath89 .
this edge may be deleted later on . deleting an edge @xmath502 from @xmath89 happens either in step 6 of the trimming rule or in step 4 of the path reduction rule . however , both rules add a new edge @xmath503 to @xmath89 that is adjacent to all the edges that were previously adjacent to @xmath502 and that were not deleted by the two rules .
hence , @xmath501 is still adjacent to an edge of @xmath89 in @xmath13 .
in other words , @xmath227 or @xmath228 is incident to an edge of @xmath89 in @xmath13 .
because @xmath227 and @xmath228 belong to @xmath489 , this is not possible due to lemma [ l - claim1 ] .
hence , @xmath498 . in order to show that @xmath499 , we again use the observation that whenever the trimming or path reduction rule deletes an edge @xmath504 , the rule adds a new edge @xmath503 in @xmath89 such that @xmath503 is adjacent to all the edges @xmath220 that were previously adjacent to @xmath502 and that were not deleted by the rules . in this case
we make the extra observation that if a vertex @xmath11 is an end - vertex of @xmath502 that is not deleted by the rule , then @xmath11 is an end - vertex of @xmath503 . because the vertices in @xmath489 are not incident to any edges in @xmath89 by lemma [ l - claim1 ] , we find that @xmath190 was incident to an edge of @xmath89 after applying the trimming rule or path reduction rule that added the edge @xmath505 to @xmath90 .
we also observe that an edge in @xmath90 is only deleted from @xmath90 if one of its end - vertices is deleted unless it is added to @xmath89 by the path reduction rule .
this means that we can argue as follows .
first suppose that @xmath505 was added to @xmath90 due to an application of the trimming rule .
if @xmath228 was adjacent to @xmath190 when the rule was applied , then @xmath506 was included in @xmath90 as well by the definition of this rule .
if @xmath228 was made adjacent to @xmath190 by the path reduction rule afterwards , then @xmath499 by the definition of the path reduction rule .
now suppose that @xmath505 was added to @xmath90 due to an application of the path reduction rule . by definition of this rule ,
@xmath227 and @xmath190 were not adjacent to each other before .
suppose that @xmath507 .
then @xmath508 are edges of the original input graph of the tree @xmath83 edges square root problem . because @xmath505 was not such an edge , @xmath227 and @xmath228 only became true twins due to an application of the path reduction rule .
then , by lemma [ l - twins ] , @xmath227 or @xmath228 must be an outer vertex of some @xmath169-triple , that is , at least one of these two vertices must be incident to an edge of @xmath89 .
then there is an edge of @xmath89 incident to at least one of these two vertices after the exhaustive application of the path reduction rule . because @xmath227 and @xmath228 are in @xmath489 , this is a contradiction to lemma [ l - claim1 ] .
hence , @xmath499 .
this completes the proof of lemma [ l - claim2 ] .
we prove the following lemma , which is our final lemma ; in particular note that if @xmath491 exists then its size is bounded by a quadratic function of @xmath0 .
[ lem : final ] if the simplicial vertex reduction rule returned a no - answer for @xmath243 , then @xmath243 has no solution . otherwise , the new instance @xmath490 has a solution if and only if @xmath243 has a solution .
moreover , @xmath491 has at most @xmath509 vertices .
we start by showing that @xmath243 has no solution if the simplicial vertex reduction rule returned a no - answer for @xmath243 .
this can happen in step 2 , 4 , 5 or 6 , each of which we discuss in a separate case .
* case 1 . *
the no - answer is given in step 2 .
suppose @xmath243 has a solution @xmath14 .
we will prove that @xmath510 , which means that returning a no - answer is correct if @xmath474 .
let @xmath236 be the set of pendant vertices of @xmath14 , and let @xmath453
. by observation [ obs : leaves ] i ) , vertices in @xmath236 are simplicial vertices of @xmath13 .
by lemma [ lem : reduced ] ii ) and iii ) , we find that @xmath511 .
hence , @xmath512 , where the last inequality follows from lemma [ l - hstarbounded ] .
* case 2 . *
the no - answer is given in step 4 .
suppose @xmath243 has a solution @xmath14 .
we will prove that @xmath513 , which means that returning a no - answer is correct if @xmath514 .
let @xmath233 be the graph obtained from @xmath14 after removing all pendant vertices of @xmath14
. then @xmath515 by lemma [ l - hstarbounded ] .
if a set @xmath472 contains a pendant vertex @xmath11 of @xmath14 , then @xmath11 is adjacent to a vertex @xmath12 of @xmath233 .
then , by observation [ obs : leaves ] ii ) , @xmath12 is not adjacent to pendant vertices of @xmath14 in any @xmath473 with @xmath516 .
otherwise @xmath472 consists of non - pendant vertices of @xmath14 , that is , vertices of @xmath233 ; being nonempty @xmath472 contains at least one vertex of @xmath233 .
we conclude that every set in the true twin partition of @xmath64 corresponds to at least one unique vertex of @xmath233 .
if their total number @xmath514 , this means that @xmath517 ; a contradiction .
hence , @xmath518 , as we had to show .
* case 3 . *
the no - answer is given in step 5 .
suppose that @xmath243 has a solution @xmath14 .
we will prove that the edges of @xmath441 incident to a set @xmath477 have a common end - vertex for @xmath484 , which means that returning a no - answer is correct should this not be the case . in order to obtain a contradiction ,
suppose that some set @xmath477 contains two vertices @xmath11 and @xmath12 that are incident to edges @xmath519 with @xmath520 .
by lemma [ lem : reduced ] ii ) , we find that @xmath521 and @xmath522 are incident to pendant vertices of @xmath14 . by observation [ obs :
leaves ] iii ) , these pendant vertices are not adjacent in @xmath13 .
however , from the definition of @xmath472 we deduce that @xmath523 are mutually adjacent ; a contradiction .
this completes case 3 .
* case 4 . *
the no - answer is given in step 6 .
then there exists a set @xmath472 such that @xmath478 and such that there are three vertices @xmath479 , @xmath480 and @xmath481 with @xmath482 and @xmath483 . in order to obtain a contradiction ,
assume that @xmath243 has a solution @xmath14 . by lemma [ l - hstarbounded ]
, @xmath14 has at most @xmath452 non - pendant vertices . because @xmath478 , this means that at least one vertex @xmath524 is a pendant vertex of @xmath14 .
also , @xmath479 is a pendant vertex of @xmath14 that has @xmath12 as its unique neighbor , because @xmath482 and all other edges incident to @xmath11 belong to @xmath90 by definition of @xmath64 . if @xmath525 , then @xmath12 is not adjacent to @xmath228 in @xmath14 , because @xmath483 . if @xmath526 , then @xmath12 is not adjacent to @xmath228 in @xmath14 either , because @xmath483 and @xmath494 ( due to lemma [ l - claim2 ] ) imply @xmath527 .
we conclude that @xmath11 and @xmath228 are pendant vertices of @xmath14 adjacent to different vertices .
however , from observation [ obs : leaves ] iii ) we derive that @xmath11 and @xmath228 are not adjacent in @xmath13 .
this is a contradiction , because @xmath11 and @xmath228 are true twins in @xmath13 by definition of @xmath472 .
this completes case 4 . from now on assume that the simplicial vertex reduction rule did not return a no - answer after performing step 6 . let @xmath528 be the instance created after applying step 7 to some set @xmath529 with @xmath39 , that is , @xmath217 is the graph obtained from @xmath13 after deleting @xmath530 , whereas @xmath257 and @xmath258 are the sets obtained from @xmath89 and @xmath90 , respectively , after deleting edges incident to @xmath530 from them .
we claim that @xmath528 has a solution if and only if @xmath243 has a solution .
before we prove this claim , we first observe that in any solution @xmath14 for @xmath243 the vertices @xmath531 are pendant vertices in @xmath14 .
this is because @xmath531 are incident to exactly one edge in @xmath441 , whereas all the other edges incident to them belong to @xmath90 .
moreover , @xmath531 have a ( unique ) common neighbor in @xmath14 , as otherwise a no - answer would have been returned in step 5 .
we let @xmath12 denote this common neighbor .
similarly , @xmath191 is a pendant vertex that has @xmath12 as its ( unique ) neighbor in any solution @xmath130 for @xmath528 .
first suppose that @xmath528 has a solution @xmath130 . then the graph obtained from @xmath130 by adding the vertices @xmath532 and the edges @xmath533 is a square root of @xmath13 by lemma [ lem : twins ] i ) . by definition of @xmath257 , @xmath258 and the set @xmath477 (
all of whose vertices are incident to one edge of @xmath534 and to edges in @xmath90 ) it is a solution for @xmath243 as well .
now suppose that @xmath243 has a solution @xmath14 .
then the graph obtained from @xmath14 after deleting @xmath532 is a square root of @xmath217 by lemma [ lem : twins ] ii ) . by definition of @xmath257 and @xmath258 ,
it is a solution for @xmath528 as well .
we denote the instance resulting from step 7 by @xmath243 again and observe that every @xmath477 now contains at most one vertex .
it remains to consider what happens at step 8 .
we let @xmath528 be the instance created after applying step 8 to some set @xmath472 with @xmath487 , that is , @xmath217 is the graph obtained from @xmath13 after deleting a set @xmath535 of @xmath536 arbitrary vertices from @xmath489 ( note that this is possible as @xmath537 ) , whereas @xmath257 and @xmath258 are the sets obtained from @xmath89 and @xmath90 , respectively , after deleting the edges that are incident to vertices of @xmath535 .
we claim that @xmath528 has a solution if and only if @xmath243 has a solution .
first suppose that @xmath528 has a solution @xmath130 .
because we could not apply the trimming and path reduction rules for @xmath243 , we can not apply these rules for @xmath528 either . then , by using the same arguments that we applied for @xmath243 in the proof of lemma [ l - hstarbounded ] , we find that @xmath130 contains at most @xmath452 non - pendant vertices .
note that @xmath130 contains at least @xmath538 vertices , which are all in @xmath472 .
hence , @xmath130 has at least one pendant vertex @xmath227 that belongs to @xmath472 .
let @xmath12 be the ( unique ) vertex adjacent to @xmath227 in @xmath130 .
then the graph @xmath14 obtained from @xmath130 by adding the vertices of @xmath535 and their edges incident to @xmath12 is a square root of @xmath13 by lemma [ lem : twins ] i ) .
we argue that @xmath14 is a solution for @xmath243 as well .
because the vertices of @xmath539 are not incident to the edges of @xmath89 due to lemma [ l - claim1 ] , we have to show that none of the @xmath540 edges that we added in order to obtain @xmath14 belong to @xmath90 .
if @xmath481 , then @xmath541 and because @xmath494 for all @xmath524 , we have that @xmath542 for all @xmath543 .
assume that @xmath544 .
recall that @xmath537 after step 7 . because @xmath487 after step 7 , @xmath478 .
then @xmath542 for all @xmath524 as otherwise the algorithm would have produced a no - answer at step 6 .
now suppose that @xmath243 has a solution @xmath14 . by lemma
[ l - hstarbounded ] , the graph @xmath14 contains at most @xmath452 non - pendant vertices . hence
, @xmath14 has at least @xmath545 pendant vertices .
because vertices in @xmath489 are true twins not incident to edges of @xmath89 and @xmath494 for any @xmath495 , we may assume without loss of generality that the vertices of @xmath535 are amongst these pendant vertices of @xmath14 .
if @xmath546 , then @xmath227 is a pendant vertex in @xmath14 incident to a unique edge @xmath547 . by observation [ obs : leaves ] , all pendant vertices of @xmath14 that are in @xmath472 are adjacent to @xmath12 in @xmath14 . then the graph obtained from @xmath14 after deleting the vertices of @xmath535 is a square root of @xmath217 by lemma [ lem : twins ] ii ) . by definition of @xmath257 and @xmath258 ,
it is a solution for @xmath528 as well .
if @xmath548 , then all pendant vertices of @xmath14 that are in @xmath472 are adjacent to some @xmath12 in @xmath14 by observation [ obs : leaves ] . then , by lemma [ lem : twins ] ( ii ) , the graph obtained from @xmath14 by deleting the vertices of @xmath535 is a square root of @xmath217 . by definition of @xmath257 and @xmath258 ,
it is a solution for @xmath528 as well . from the above
it follows that the instance @xmath490 obtained after step 8 has a solution if and only if @xmath243 has a solution . in order to complete the proof
, we must show that @xmath491 has at most @xmath509 vertices .
each @xmath472 has at most @xmath538 vertices due to step 8 , and we also have @xmath518 due to step 4 .
hence @xmath549 . as
the number of vertices in @xmath550 is at most @xmath452 due to step 2 , we obtain that @xmath551 , as required .
let @xmath1 and @xmath2 denote the number of vertices and edges of the graph @xmath13 of the original instance @xmath242 of tree @xmath83 edges square root . in order to complete the proof of theorem [ thm : tree - few - edges ] , we first note that the trimming and path reduction rules are applied at most @xmath1 times to construct the instance @xmath490 .
each application of the trimming rule can be done in time @xmath552 and each application of the path reduction rule takes time @xmath553 .
finally , the simplicial vertex reduction rule can be done in time @xmath554 .
hence , our kernelization algorithm runs in time @xmath555 , and it remains to solve the obtained reduced instance @xmath490 . because @xmath491 has at most @xmath509 vertices , @xmath491 has at most @xmath556 edges .
therefore , we can solve tree @xmath83 edges square root with labels for instance @xmath490 in time @xmath557 ; we consider all edge subsets of @xmath491 that have size at most @xmath558 and use brute force .
we conclude that the total running time of our algorithm is @xmath559 , as required .
we finish this section with the following remarks .
first , recall that our quadratic kernel is a generalized kernel for the tree @xmath83 edges square root problem .
we believe that a quadratic kernel exists for this problem as well by using a similar reduction .
however , proving this seemed to be more technical and also to yield a graph with more than @xmath509 vertices .
we therefore chose to prove our fpt result by using a reduction leading to a generalized kernel .
second , it should also be noted that our generalized kernel for tree @xmath83 edges square root does not imply a kernel for tree @xmath83 edges square root with labels , because our reduction rules require that the original instance is unlabeled .
we do not know whether the ( more general ) problem tree @xmath83 edges square root with labels is fpt as well .
recall that the maximum square root problem is that of testing whether a given graph @xmath13 with @xmath2 edges has a square root with at least @xmath33 edges for some given integer @xmath33 . in this section
we give an fpt algorithm for this problem with parameter @xmath43 .
in other words , we show that the problem of deciding whether a graph @xmath13 has a square root that can be obtained by removing at most @xmath0 edges of @xmath13 is fixed - parameter tractable when parameterized by @xmath0 .
we also present an exact algorithm for the maximum square root problem .
both algorithms are based on the observation that in order to construct a square root @xmath14 from a given graph @xmath13 , we must delete at least one of every pair of adjacent edges that do not belong to a triangle in @xmath13 .
we therefore construct an auxiliary graph @xmath560 that has vertex set @xmath44 and an edge between two vertices @xmath561 and @xmath562 if and only if @xmath563 and @xmath564 for three distinct vertices @xmath565 with @xmath566 .
observe that @xmath560 is a spanning subgraph of the line graph of @xmath13 .
we need the following lemma .
[ lem : charact ] let @xmath14 be a spanning subgraph of a graph @xmath13 . then @xmath14 is a square root of @xmath13 if and only if @xmath347 is an independent set of @xmath560 and every two adjacent vertices in @xmath13 are at distance at most @xmath66 in @xmath14 .
first suppose that @xmath14 is a square root of @xmath13 . by definition ,
every two adjacent vertices in @xmath13 are of distance at most 2 in @xmath14 . in order to show that @xmath347 is an independent set in @xmath560 ,
assume that two edges @xmath567 are adjacent vertices in @xmath560
. then @xmath563 and @xmath564 for three distinct vertices @xmath565 with @xmath566 .
this means that @xmath227 and @xmath190 are of distance 2 in @xmath14 implying that @xmath568 , which is a contradiction .
now suppose that @xmath347 is an independent set of @xmath560 and that every two adjacent vertices in @xmath13 are at distance at most @xmath66 in @xmath14 . in order to show that @xmath14 is a square root of @xmath13 , it suffices to show that every two non - adjacent vertices in @xmath13 have distance at least 3 in @xmath14 .
let @xmath11 and @xmath12 be two non - adjacent vertices in @xmath13 that have distance at most 2 in @xmath14 .
then there exists a vertex @xmath569 such that @xmath570 . then @xmath571 and @xmath572 are adjacent in @xmath560 contradicting the independence of @xmath347 in @xmath560 .
we use lemma [ lem : charact ] to prove propositions [ p - fpt ] and [ p - exact ] . here
, we use the @xmath573-notation to suppress any polynomial factors .
vertex cover _ is a subset @xmath574 such that every edge is incident with at least one vertex of @xmath48 .
the vertex cover problem is that of testing whether a given graph has a vertex cover of size at most @xmath575 for a given integer @xmath575 . in proposition [ p - fpt ]
we prove that there is a @xmath576 time algorithm to decide whether a given graph @xmath13 has square root @xmath14 such that @xmath577 .
[ p - fpt ] maximum square root can be solved in time @xmath576 .
let @xmath13 be a graph with @xmath1 vertices and @xmath2 edges , and let @xmath5 be an integer . by lemma [ lem : charact ] it suffices to check whether @xmath560 has a vertex cover @xmath48 of size at most @xmath0 such that @xmath578 is a square root of @xmath13 .
all vertex covers of size at most @xmath0 of a graph can be enumerated by adapting the standard @xmath576 branching algorithm for the vertex cover problem ( see for example @xcite ) .
it requires @xmath579 time to compute @xmath560 and @xmath554 time to check whether a graph @xmath580 is a square root of @xmath13 .
hence the overall running time of our algorithm is @xmath576 .
we observe that maximum square root has a linear kernel for connected graphs .
this immediately follows from a result of aingworth , motwani and harary @xcite , who proved that if @xmath14 is a square root of a connected @xmath1-vertex graph @xmath581 , then @xmath582 .
hence , @xmath583 for every yes - instance @xmath242 of maximum square root with @xmath581 ( trivially , @xmath71 is its own square root ) .
note that this kernel does not lead to a faster running time than @xmath576 .
in proposition [ p - exact ] we present our exact algorithm , which does not only solve the decision problem but in fact determines a square root of a given graph that has maximum number of edges .
[ p - exact ] maximum square root can be solved in time @xmath584 on graphs with @xmath2 vertices .
let @xmath13 be a graph with @xmath1 vertices and @xmath2 edges , and let @xmath5 be an integer .
we compute the graph @xmath560 , enumerate all maximal independent sets @xmath585 of @xmath560 , and verify for each @xmath586 whether @xmath13 is the square of the graph @xmath587 . out of those graphs
@xmath588 that are square roots of @xmath13 , return the one with maximum number edges ; if no such graph @xmath588 has been found , then @xmath13 has no square roots .
correctness follows from lemma [ lem : charact ] .
recall that @xmath560 can be computed in time @xmath579 .
all the maximal independent sets of the @xmath2-vertex graph @xmath560 can be enumerated in time @xmath584 using the polynomial delay algorithm of tsukiyama et al .
@xcite , since @xmath560 has at most @xmath589 maximal independent sets @xcite .
finally , recall that for each maximal independent set @xmath585 , we can check in time @xmath554 whether @xmath590 .
hence the overall running time of our algorithm is @xmath584 .
we conclude our paper with two open problems .
first , is it also possible to construct an exact algorithm for minimum square root that is better than the trivial exact algorithm ?
second , recall that if @xmath14 is a square root of a connected @xmath1-vertex graph @xmath581 , then @xmath582 @xcite .
is it fpt to decide whether a connected @xmath1-vertex graph @xmath581 has a square root that can be obtained by removing at most @xmath591 edges , or equivalently , whether a connected @xmath1-vertex graph @xmath581 has a square root with at least @xmath592 edges , when parameterized by @xmath0 ?
in particular , can it be decided in polynomial time whether a connected graph @xmath13 has a square root with _ exactly _
@xmath593 edges ? | we show that the following two problems are fixed - parameter tractable with parameter @xmath0 : testing whether a connected @xmath1-vertex graph with @xmath2 edges has a square root with at most @xmath3 edges and testing whether such a graph has a square root with at least @xmath4 edges .
our first result implies that squares of graphs obtained from trees by adding at most @xmath0 edges can be recognized in polynomial time for every fixed @xmath5 ; previously this result was known only for @xmath6 .
our second result is equivalent to stating that deciding whether a graph can be modified into a square root of itself by at most @xmath0 edge deletions is fixed - parameter tractable with parameter @xmath0 . |
topological superconductor ( ts ) has received considerable experimental and theoretical attentions because majorana zero - energy modes appear at the ends of the one - dimensional ts which can potentially be used for decoherence - free quantum computation.@xcite in comparison with the conventional superconductor , the ts system shows new and interesting properties.@xcite for instance , in the proximity - coupled semiconductor - ts devices , the majorana zero modes induce the zero - bias anomaly.@xcite a more compelling ts signature is the unusual josephson current - phase relation .
namely , when the normal s - wave superconductor nano - wire is replaced by a ts wire with the majorana zero modes , the current - phase relation will be modified to be @xmath1 and the period of the josephson current vs @xmath2 will be @xmath3 ( @xmath2 is the superconducting phase difference ) .
this is the so - called the fractional josephson effect.@xcite such a result can be understood in terms of fermion parity ( fp ) .
if the fp is preserved , there will be a protected crossing of the majorana bound states at @xmath4 with perfect population inversion . as a result
, the system can not remain in the ground state as @xmath2 evolves from @xmath5 to @xmath6 adiabatically.@xcite recently , a new class of topological superconductors with time - reversal symmetry , referred to as a diii symmetry - class superconductor and classified by the @xmath7 topological invariant,@xcite has attracted rapidly growing efforts.@xcite differently from chiral superconductors , in diii - class superconductors , the zero modes come in pairs due to kramers s theorem . up to now , many schemes have been proposed to realize @xmath7 time - reversal - invariant majorana nanowires using the proximity effects of @xmath8-wave , @xmath9-wave , @xmath10-wave , or conventional @xmath11-wave superconductors.@xcite it was shown that at each end of such a nanowire are localized two majorana bound states that form a kramers doublet and are protected by time - reversal symmetry .
it is certain that the majorana doublet can induce some new and interesting transport behaviors .
some groups have reported that in the josephson junction formed by the majorana doublet , the josephson currents show different periods in the cases of different fps.@xcite however , for completely describing the transport properties of the majorana doublet , any new proposals are desirable . in this work ,
we aim to investigate the influence of an embedded quantum dot ( qd ) on the current properties in the josephson junction contributed by the majorana doublets .
our motivation is based on the following two aspects .
first , qd is able to accommodate an electron and the average electron occupation in one qd can be changed via shifting the qd level .
thus , when one qd is introduced in the ts junction , the fp can be re - regulated and the fractional josephson current can be modified .
moreover , some special qd geometries can induce the typical quantum interference mechanisms , e.g. , the fano interference,@xcite which are certain to play an important role in adjusting the fractional josephson effect .
second , one qd can mimic a quantum impurity in the practical system , which is able to provide some useful information for relevant experiments .
our calculations show that when the majorana doublets couple indirectly via a non - magnetic qd , only the normal josephson effects take place , irrelevant to the change of the fp .
when a finite magnetic field is applied on the qd , the fractional josephson effect comes into being in the odd - fp case . on the other hand ,
when the direct and indirect couplings between the majorana doublets coexist , no fractional josepshon effect occurs despite the presence of magnetic field on the qd , and the periods of the josephson currents have opportunities to present the @xmath0-to-@xmath6 transition when the qd level is shifted .
the results in this work will be helpful for describing the qd - assisted josephson effect between the majorana doublets .
the josephson junction that we consider is formed by the direct couplings between the majorana nanowires and their indirect couplings via a qd , as illustrated in fig.1 .
the particle tunneling process in this junction can be described by hamiltonian @xmath12 with @xmath13 @xmath14 denotes the particle motion in the two majorana nanowires . with the proximity - induced @xmath9-wave and @xmath11-wave superconducting pairs , the effective tight - binding hamiltonian in the @xmath15-th nanowire
can be written as@xcite @xmath16 @xmath17 and @xmath18 ( @xmath19 or @xmath20 ) are the electron creation and annihilation operators for the @xmath21-th site in the @xmath15-th nanowire .
@xmath22 is the inter - site hopping energy and @xmath23 represents the strength of spin - orbit coupling . @xmath24 and @xmath25 denote the energies of the @xmath9-wave and @xmath11-wave superconducting pairings , respectively .
@xmath26 is the chemical potential in the @xmath15-th nanowire .
note that the hopping coefficients and the chemical potential are generically reonormalized by the proximity effect . the second term @xmath27 denotes the direct coupling between the two majorana nanowires , which can be expressed as @xmath28 where @xmath29 is the direct coupling coefficient .
next , @xmath30 is to express the indirect coupling between the two majorana nanowires due to the presence of an embedded qd ( or a quantum impurity ) .
its expression can be given by @xmath31 here @xmath32 and @xmath33 are the electron creation and annihilation operators in the qd , and @xmath34 is the qd level .
@xmath35 denotes the strength of an effective magnetic field applied on the qd , and @xmath36 denotes the intradot electron interaction with @xmath37 .
in addition , @xmath38 is coupling coefficient between the qd and the @xmath15-th majorana nanowire . in order to discuss the josephson effect in this junction
, we have to deduce an effective hamiltonian that reflects the direct and indirect couplings between the majorana doublets . for this purpose
, we define the majorana operators @xmath39 , \notag\\ & & \tilde{\gamma}_{\alpha1}=\sum_j[\tilde{\mu}^{(1)}_{\alpha j}\tilde{c}_{\alpha j}+\tilde{\mu}^{(1)*}_{\alpha j}\tilde{c}^\dag_{\alpha j } ] , \notag\\ & & \gamma_{\alpha2}=\sum_j[\mu^{(2)}_{\alpha j}c_{lj}+\mu^{(2)*}_{\alpha j}c^\dag_{\alpha j}],\notag\\ & & \tilde{\gamma}_{\alpha2}=\sum_j[\tilde{\mu}^{(2)}_{\alpha j}\tilde{c}_{\alpha j}+\tilde{\mu}^{(2)*}_{\alpha j}\tilde{c}^\dag_{\alpha j}],\end{aligned}\ ] ] where @xmath40 is the renormalized electron operator at the @xmath21-th site in the @xmath15-th site with @xmath41 . using the above formulas
, we can solve the electron operators in terms of majorana and nonzero - energy quasiparticle operators .
reexpressing the quasiparticles in terms of electron operators , we can interpret @xmath42 , @xmath43 , @xmath44 , and @xmath45 by @xmath46 in which the normalization factor has been neglected .
besides , @xmath47 , @xmath48 and @xmath49 , @xmath50 are expansion coefficients , originated from the quasiparticle operators other than the corresponding majorana mode . substituting eq.([transform ] ) into the expression of @xmath12 , we can obtain the low - energy effective hamiltonian of @xmath12 in the case of infinitely - long nanowires , which is divided into two parts .
the first part is @xmath51 the relevant parameters here are defined as follows : @xmath52 , @xmath53 , and @xmath54 in which @xmath55 , @xmath56 , @xmath57 , and @xmath58 ( it is reasonable to suppose @xmath59 and @xmath60 ) . besides , in the above formula @xmath61 , @xmath62 , and @xmath63 with @xmath64 and @xmath65 . for the second part ,
when the highest - order terms are neglected , it can be approximated as @xmath66 it should be noted that since the @xmath11-wave pairing is present in the quantum wires , the electrons @xmath67 and @xmath68 will form a cooper pair and condense .
this process leads to an effective coupling between majorana zero modes localized at the same end and the finite coupling between the kramers doublet in the qd .
therefore , up to the second - order perturbation in the tunneling process , we can express @xmath69 as @xmath70 where @xmath71 with @xmath72 being a time - ordered integral . in the case of uniform superconducting pairings in the majorana nanowires ,
@xmath73 can be further deduced as @xmath74 in which @xmath75 is the eigen - energy of the isolated superconductor.@xcite with the relations in eq.([transform ] ) , we can get the relationship that @xmath76 and @xmath77 .
accordingly , @xmath69 can be written as @xmath78 in which @xmath79 , @xmath80 , and @xmath81 .
up to now , we have obtained the low - energy effective hamiltonian of such a structure .
the phase difference between the two majorana wires will drive finite josephson current through them , which can be directly evaluated by the following formula @xmath82 with @xmath83 being the thermal average .
it is certain that solving the josephson current is dependent on the diagonalization of @xmath84 . in the following ,
we try to diagonalize the hamiltonian .
first , by defining @xmath85 and @xmath86 with @xmath87 , we reexpress @xmath12 as @xmath88 where @xmath89 is supposed to be @xmath90 .
next , @xmath84 can be expressed in the normal fermion representation by supposing @xmath91 , @xmath92 and @xmath93 , @xmath94 where @xmath95 , @xmath96 and @xmath97 , @xmath98 are the fermionic creation and annihilation operators .
accordingly , the matrix form of @xmath84 can be deduced on the basis of @xmath99 where @xmath100 and @xmath101 .
note that in the system with majorana bound states , only the fp is the good quantum number .
thus , we should build the fock state according to the fp .
first , in the case of the even fp , the fock state can be written as @xmath102 . as a result
, the matrix form of @xmath103 can be written as @xmath104,\label{matrix1}\end{aligned}\ ] ] where @xmath105 , @xmath106 , @xmath107 , @xmath108 , and @xmath109 .
next , for the case of the even fp , the fock state can be written as @xmath110 and the matrix of @xmath111 takes the form as @xmath112.\label{matrix2}\end{aligned}\ ] ] for the extreme case of strong magnetic field limit , if @xmath113 is in the finite - energy region , @xmath114 will be empty , and then only one level contributes to the josephson effects , respectively .
accordingly , in such a case , the matrixes of @xmath115 and @xmath116 will be halved , i.e. , @xmath117,\label{matrix5}\end{aligned}\ ] ] and @xmath118.\label{matrix6}\end{aligned}\ ] ] based on the above analysis , we can understand that at the zero - temperature limit , the josephson current in this structure is dependent on the fp , i.e. , @xmath119 @xmath120 are the ground - state ( gs ) energies in the even and odd fp cases , respectively .
following the derivation in the above section , we next perform the numerical calculation to discuss the detailed properties of the josephson current through such a system . as a typical case ,
only the zero - temperature limit is considered . before calculation
, we would like to review the josephson effect in the case of @xmath121 .
in such a case , @xmath122 , and @xmath123 are the eigenstates of @xmath84 .
the two even - fp eigenstates are @xmath124 and @xmath125 , and their corresponding gs energies are @xmath126 .
contrarily , the odd - fp eigenstates are @xmath127 and @xmath128 with the gs energies @xmath129 . just as concluded in the previous works,@xcite the fractional josephson effect occurs in the situation of even fp , otherwise only the normal josephon effect can be observed . in fig.2
we suppose @xmath130 and choose @xmath131 and @xmath132 to investigate the josephson effect in the case where the majorana doublets couple indirectly to each other via a qd .
the results are shown in figs.2 - 3 : fig.2 corresponds to the noninteracting results , and fig.3 describes the influences of the intradot coulomb interaction on the josephson effects in different fps in the case of @xmath133 .
first , in fig.2(a)-(b ) we find that when a non - magnetic qd is presented , it induces the occurrence of the normal josephson effects and the departure of @xmath34 from zero weakens the current amplitudes , irrelevant to the fp difference .
also , the fp plays an important role in affecting the josephson effects .
concretely , the josephson currents in different fps flow in the opposite directions for the same @xmath2 , and the amplitude of @xmath134 is about one half of that of @xmath135 and when @xmath136 @xmath134 gets close to zero .
besides , at the points of @xmath137 , in the even - fp case the discontinuous change of the josephson current is more well - defined compared with the odd - fp case .
next , when finite magnetic field is applied on the qd , the even - fp josephson current shows little change except that its amplitude becomes less dependent on the qd - level shift .
however , in the odd - fp case , the josephson current changes completely . it can be clearly found that with the strengthening of the magnetic field , the original current oscillation is suppressed . especially in the vicinity of @xmath138 , the current amplitude tends to disappear at the case of @xmath139 .
thus , it is certain that in the case of odd fp , a nonzero field on the qd can induce the occurrence of the fractional josephson current .
in addition , the increase of @xmath35 enhances the current oscillation around the points of @xmath137 when @xmath34 departs from @xmath140 . up to now
, we can conclude that when the majorana doublets are coupled by a magnetic qd , the fractional josephson effect has an opportunity to take place , but it is different from the case where the majorana doublets couple directly.@xcite coulomb interaction is a key factor to influence the characteristics of qd .
in fig.3 we consider the case of finite intradot coulomb interaction and investigate the effect of the magnetic qd on the josephson currents in the finite - coulomb case . in fig.3(a)-(b ) we first find that in the even - fp case with a non - magnetic qd , the coulomb interaction benefits the josephson effect , since in the region of @xmath141 the current amplitude is relatively robust and weakly dependent on the shift of the qd level . in contrast , for the odd - fp case , the intradot coulomb interaction only moves the current maximum to the point of @xmath142 , but it does not vary the current oscillation manner compared with the noninteracting case .
hence , the coulomb interaction only adjusts the effect of the qd level on the josephson effects but does not modify the current oscillation manner with the change of @xmath2 .
next , fig.3(c)-(h ) show that regardless of the fp difference , the current amplitudes are suppressed by the application of finite magnetic field on the qd . in the even - fp case ,
the current amplitude around the point of @xmath142 undergoes a relatively - apparent suppression . for the odd - fp case , except the suppression of the current amplitude ,
the fractional josephson effect becomes weak but can still be observed . according to the results in the above paragraphes ,
when the majorana doublets couple indirectly via a magnetic qd , the fractional josephson effect has an opportunity to come into being in the odd - fp case . otherwise , only the normal josephson effect can be observed . in order to explain the results in figs.2 - 3
, we would like to compare the case of @xmath143 and @xmath121 with the case of @xmath144 and @xmath145 . in the former case ,
the josephson effects are only determined by the fp of state @xmath123 .
namely , when the system is located at the states @xmath124 or @xmath125 , the fractional josephson effect occurs .
however , when the majorana doublets couple indirectly via a qd , the fock space defined by @xmath123 just becomes a subspace of the fock space formed by @xmath99 .
since both @xmath146 and @xmath90 are equal to zero in such a case , the nonzero contributions of @xmath124 and @xmath125 to the josephson effect are dependent on their finite indirect couplings , i.e. , their simultaneous couplings to the other states .
thus , the manners of the indirect couplings inevitably regulate the properties of the josephson effects .
for instance , in eq.([matrix1 ] ) we find that in the even - fp case , the states that include @xmath124 and @xmath125 couple indirectly in the way of left - right symmetry .
this leads to the equal contributions of @xmath124 and @xmath125 to the josepshon effect , and then the normal josephon effect takes place , independent of the presence of magnetic field . on the other hand , in the odd - fp case , eq.([matrix1 ] )
shows that the coupling between the states that include @xmath124 and @xmath125 is left - right asymmetric unless @xmath147 .
consequently , if finite @xmath35 is considered , @xmath124 and @xmath125 make different contributions to the josephson effect , which gives rise to the fractional josephson effect .
also , note that in odd - fp case , the left - right asymmetric coupling manner weakens the quantum coherence and suppresses the current amplitude to some degree .
when the coulomb interaction is taken into account , the qd level splits . in the even - fp case ,
the picture in the noninteracting case can be doubled since the left - right symmetric coupling manner .
nevertheless , in the odd - fp case , only in the case of electron - hole symmetry , the left - right symmetric coupling manner can be satisfied .
therefore , the maximal josephson current occurs at the point of @xmath142 in the case of @xmath133 .
we next proceed to pay attention to the josephson effect in the case where the direct and indirect couplings between the majorana doublets coexist .
the results are shown in figs.4 - 5 where @xmath146 is taken to be @xmath148 and @xmath149 .
the noninteracting results are presented in fig.4 and fig.5 describes the case of @xmath133 . in fig.4
, we find that for any @xmath34 , the two kinds of josephson currents show dissimilar oscillations with the adjustment of superconducting phase difference . in the even - fp case with @xmath150 , when @xmath34 gets approximately close to @xmath151 , the amplitude of the josephson current is decreased , otherwise , the josephson effect will be enhanced and then holds .
however , the current period always keeps to be @xmath6 , as shown in fig.4(a ) .
next , in fig.4(b ) where @xmath152 , we can see that the only effect of the magnetic field in the qd is to further suppress the minimum of the josephson current .
such a result is exactly similar to the result of @xmath153 . on the other hand , for the odd - fp case , in fig.4(c ) , it shows that in the region of @xmath154 , the josephson current seems to oscillate in @xmath0 period . when the qd level increases to @xmath142 , the current period experiences the @xmath0-to-@xmath6 transition with the disappearance of the current around the points of @xmath155
. next , in the region of @xmath156 , the josephson current varies in @xmath6 period with its maximum in the vicinity of @xmath157 .
when @xmath34 further increases from @xmath158 , the josephson current recovers the @xmath0-period oscillation gradually .
fig.4(d ) presents the effect of the magnetic field on the qd in odd - fp case .
it seems that in such a case , the magnetic field can not induce the fractional josephson effect , but it tends to enhance the current amplitude in the region of @xmath156 , which is exactly opposite to the case of @xmath153 . following the above result
, we present the influence of the magnetic field on the finite - coulomb results , as displayed in fig.5 . here
, the coulomb strength is also taken to be @xmath133 .
first , fig.5(a ) shows the even - fp result with the non - magnetic qd .
we can find that in such a case , the current minimum is shifted to the position of @xmath159 . besides , the coulomb interaction efficiently weakens the josephson effect , since increasing @xmath34 from @xmath160 begins to eliminate the current amplitude gradually .
next when finite magnetic field is applied on the qd with @xmath152 , it further suppresses the minimum of the josephson current , similar to the noninteracting case [ see fig.5(b ) ] .
the odd - fp results are shown in fig.5(c)-(d ) , where the magnetic field strength is taken to be zero and @xmath161 , respectively . in fig.5(c ) , we see that different from the noninteracting result , the @xmath6-period oscillation of the current occurs from @xmath162 . in the region of @xmath163 ,
the josephson current varies in @xmath6 period with its maximum in the vicinity of @xmath142 . besides
, it can be noted that the coulomb interaction enhance the amplitude of the josephson current , in comparison with the noninteracting current results . for the effect of the magnetic field in the odd - fp case , as shown in fig.5(d ) , it is analogous to that in the noninteracting case .
namely , it tends to enhance the amplitude of the @xmath6-period current , despite in the region of @xmath163 .
also , it does not induce the appearance of the fractional josephson effect .
the results in figs.4 - 5 can be explained similar to the discussion about figs.2 - 3 . in the case of finite @xmath29 ,
the underlying physics that governs the josephson effects becomes complicated .
the reason arises from two aspects .
first , the fermion number of the qd re - regulates the fp of @xmath123 for conserving the fp in the whole system .
second , the fano interference can be induced by the direct and indirect couplings between the majorana doublets .
based on this idea , we see that in the even - fp case with zero magnetic field , when the qd level is far away from the energy zero point , both @xmath164 and @xmath165 are equal to @xmath166 or @xmath5 simultaneously .
consequently , the states @xmath124 and @xmath125 contribute equally to the josephson effect , so the normal josephson effect occurs with the current amplitude proportional to @xmath146 .
alternatively , in the odd - fp case with @xmath167 , the @xmath0-period current occurs and its amplitude is related to @xmath90 , due to the co - contribution of states @xmath127 and @xmath128 . when the qd level gets close to the energy zero point , it will be partly - occupied .
in such a situation , @xmath127 and @xmath128 contribute to the even - fp josephson current , whereas @xmath124 and @xmath125 devote themselves to the odd - fp current .
however , due to @xmath168 , the suppression of @xmath169 only appears in a narrow region near the point of @xmath157 , while the @xmath6-period oscillation of @xmath170 distributes in a wide region accompanied by its enhanced amplitude [ see fig.4(a ) and fig.4(c ) ] .
next , in the presence of the intradot coulomb interaction , @xmath171 splits into two , i.e. , @xmath171 and @xmath172 .
accordingly , in the energy region of @xmath173 , the fermion in the qd is changeable between 0 and 1 , which magnifies the transformation of the josephson effect caused by the shift of qd level . since a finite magnetic field on
the qd plays a similar role in affecting the fermion occupation in the qd , the magnetic field makes a similar contribution to the josephson effect , compared with the coulomb interaction .
in addition , it should be noticed that the fano interference induces the asymmetric spectra of the josepshon currents vs @xmath34 . at last
, we would like to pay attention to the extreme case of strong magnetic field where only one level ( i.e. , @xmath113 ) contributes to the josephson effects . in such a case ,
the matrix dimension of @xmath115 and @xmath116 will be halved , as discussed in the above section .
the corresponding numerical results are shown in fig.6 .
first , in fig.6(a)-(b ) we can see that in the case of @xmath153 , the josephson currents in different fps are the same as each other , with their @xmath6 period . on the other hand ,
when the direct coupling between the majorana doublets is considered , the josephson currents become dependent on the fp . as shown in fig.6(c ) , in the even - fp case , increasing @xmath174 can change the current period from @xmath0 to @xmath6 with the clear transition region near @xmath175 .
however , in the odd - fp case , similar result occurs when @xmath174 decreases .
these results can certainly be clarified by discussing the influence of the fermion number in the qd on the fp of states @xmath123 . also , in fig.6(c)-(d ) we can find that the fano interference induces the dissimilar transition behaviors of the josephson currents in different fps .
in summary , we have investigated the josephson effect in the majorana - doublet - contributed junction modified by the different inter - doublet coupling manners .
it has been found that an embedded qd in this junction plays a nontrivial role in modifying the josephson effects , since the tunable fermion occupation in the qd re - regulates the fp of the majorana doublets for conserving the fp in whole system . as a result , the @xmath3-period , @xmath6-period , and @xmath0-period josephson currents have opportunities to come into being , respectively . to be concrete , when the majorana doublets couple indirectly via a non - magnetic qd ,
the normal josephson effects occur , and the fp change just causes the reversal of the current direction and the variation of the current amplitude . only in the odd - fp case , can applying finite magnetic field on the qd induce the appearance of the fractional josephson effect .
when the direct and indirect couplings between the majorana doublets coexist , no fractional josephson effect takes place , regardless of finite magnetic field on the qd .
moreover , the @xmath0-period current has an opportunity to appear with the shift of the qd level .
we believe that this work will be helpful for describing the qd - assisted josephson effects between the majorana doublets .
this work was financially supported by the fundamental research funds for the central universities ( grants no . n130405009 and no .
n130505001 ) , the natural science foundation of liaoning province of china ( grant no .
2013020030 ) , and the liaoning baiqianwan talents program ( grant
2012921078 ) .
99 m. z. hasan and c. l. kane , rev .
phys . * 82 * , 3045(2010 ) ; x .-
qi and s .- c .
zhang , rev .
phys . * 83 * , 1057 ( 2011 ) . c. nayak , s. h. simon , a. stern , m. freedman , s. das sarma , rev .
mod . phys . * 80 * , 1083 ( 2008 ) .
f. zhang , c. l. kane , and e. j. mele , phys .
lett . * 111 * , 056402 ( 2013 ) .
f. zhang , c. l. kane , and e. j. mele , phys .
lett . * 111 * , 056403 ( 2013 ) .
f. hassler , a. r. akhmerov , c .- y .
hou , and c. w. j. beenakker , new j. phys .
* 12 * , 125002 ( 2010 ) ; k. flensberg , phys . rev .
b * 82 * , 180516(r ) ( 2010 ) .
d. pekker , c. y. hou , v. e. manucharyan , and e. demler , phys .
lett . * 111 * , 107007 ( 2013 ) ; l. jiang , d. pekker , j. alicea , g. refael , y. oreg , and f. von oppen , phys .
* 107 * , 236401 ( 2011 ) .
a. p. schnyder , s. ryu , a. furusaki , and a. w. w. ludwig , phys .
b * 78 * , 195125 ( 2008 ) .
qi , t. l. hughes , s. raghu , and s .- c .
zhang , phys .
lett . * 102 * , 187001 ( 2009 ) .
j. c. y. teo and c. l. kane , phys .
b * 82 * , 115120 ( 2010 ) .
a. p. schnyder , p.m. r. brydon , d. manske , and c. timm , phys .
b * 82 * , 184508 ( 2010 ) . c. w. j. beenakker , j. p. dahlhaus , m. wimmer , and a. r. akhmerov , phys . rev .
b * 83 * , 085413 ( 2011 ) .
s. deng , l. viola , and g. ortiz , phys .
lett . * 108 * , 036803 ( 2012 ) .
s. nakosai , y. tanaka , and n. nagaosa , phys .
lett . * 108 * , 147003 ( 2012 ) . c. l. m. wong and k. t. law , phys .
b * 86 * , 184516 ( 2012 ) . f. zhang , c. l. kane , and e. j. mele , phys .
rev . lett . * 111 * , 056402 ( 2013 ) .
s. nakosai , j. c. budich , y. tanaka , b. trauzettel , and n. nagaosa , phys .
* 110 * , 117002 ( 2013 ) .
a. keselman , l. fu , a. stern , and e. berg , phys .
111 * , 116402 ( 2013 ) .
e. gaidamauskas , j. paaske , and k. flensberg , phys .
* 112 * , 126402 ( 2014 ) .
a. haim , a. keselman , e. berg , and y. oreg , phys .
b * 89 * , 220504(r ) ( 2014 ) .
f. zhang and c. l. kane , phys .
b * 90 * , 020501(r ) ( 2014 ) .
e. dumitrescu , j. d. sau , and s. tewari , arxiv:1310.7938 .
j. klinovaja and d. loss , phys .
b * 90 * , 045118 ( 2014 ) . | one majorana doublet can be realized at each end of the time - reversal - invariant majorana nanowires .
we investigate the josephson effect in the majorana - doublet - presented junction modified by different inter - doublet coupling manners .
it is found that when the majorana doublets couple indirectly via a non - magnetic quantum dot , only the normal josephson effects occur , and the fermion parity in the system just affects the current direction and amplitude .
however , in the odd - parity case , applying finite magnetic field on the quantum dot can induce the appearance of the fractional josephson effect .
next , when the direct and indirect couplings between the majorana doublets coexist , no fractional josephson effect takes place , regardless of finite magnetic field on the quantum dot .
instead , the @xmath0-period current has an opportunity to appear in some special cases .
all the results are clarified by analyzing the influence of the fermion occupation in the quantum dot on the parity conservation in the whole system .
we ascertain that this work will be helpful for describing the dot - assisted josephson effect between the majorana doublets . |
in this paper we address the following problems ( cf . , for example , ( * ? ? ? * problem 3.2 , p. 125 and
problem 7.3 , p. 289 ) ) .
[ pr1 ] let @xmath0 .
assume that @xmath13 and @xmath14 are convex bodies in @xmath15 such that the projections @xmath16 and @xmath17 are congruent for all @xmath18 .
is @xmath13 a translate of @xmath19 ?
[ pr2 ] let @xmath0 .
assume @xmath13 and @xmath14 are star bodies in @xmath15 such that the sections @xmath20 and @xmath21 are congruent for all @xmath18 .
is @xmath22 ? here
we say that @xmath16 , the projection of @xmath13 onto @xmath7 , is congruent to @xmath17 if there exists an orthogonal transformation @xmath23 in @xmath7 such that @xmath24 is a translate of @xmath17 ; @xmath25 stands for the grassmann manifold of all @xmath6-dimensional subspaces in @xmath3 .
golubyatnikov has obtained some partial answers to problem [ pr1 ] in the case of _ direct _ rigid motions , i.e. , when the orthogonal group @xmath26 is replaced by the special orthogonal group @xmath27 . in particular
, he proved that the answer is affirmative , provided @xmath28 and none of the projections of the bodies have symmetries with respect to rotations ( in other words , the only direct rigid motion taking @xmath16 , @xmath17 onto themselves is the identity ) .
we refer the reader to @xcite , @xcite ( pp .
@xmath29 ) , @xcite ( pp .
@xmath30 ) , @xcite , @xcite , @xcite and @xcite , for history and partial results related to the above problems . in this paper
we give an affirmative answer to problems [ pr1 ] and [ pr2 ] in the class of polytopes .
our first result is [ th1 ] let @xmath0 and let @xmath1 and @xmath2 be two convex polytopes in @xmath3 such that their projections @xmath4 , @xmath5 , onto every @xmath6-dimensional subspace @xmath7 , are congruent .
then there exists @xmath31 such that @xmath32 or @xmath33 .
if projections of the bodies are _ directly _ congruent , golubyatnikov proved theorem [ th1 ] in the case @xmath28 ( the result is contained in the proof of theorem 2.1.2 , p. 19
@xcite , p. 17
, lemma 2.1.4 ) , and in the case @xmath34 under the additional assumption that none of the @xmath35-dimensional projections have rigid motion symmetries ( cf .
@xcite , p. 48 , theorem 3.2.1 ) ; see also @xcite , where the assumption on symmetries was relaxed .
our second result is [ th2 ] let @xmath0 and let @xmath1 and @xmath2 be two convex polytopes in @xmath3 containing the origin in their interior .
assume that their sections , @xmath11 , @xmath12 , by every @xmath6-dimensional subspace @xmath7 , are congruent .
then @xmath9 or @xmath10 .
if sections of the bodies are _ directly _ congruent theorem [ th2 ] is known in the case @xmath28 ( cf .
@xcite ) , and in the case @xmath34 ( cf .
@xcite ) under the additional assumptions that diameters of one of the polytopes contain the origin and certain sections related to the diameters have no rigid motion symmetries .
we remark that in both theorems it is enough to assume that only one of the bodies is a convex polytope .
this follows from a result of klee @xcite , who showed that a set in @xmath36 must be a convex polytope , provided all of its projections on @xmath6-dimensional subspaces are convex polytopes .
it follows by duality ( see @xcite , formula ( 0.38 ) , p. 22 ) that if all sections of a convex body containing the origin in its interior are polytopes , then the body is a polytope .
the paper is organized as follows . in the next section
we recall some definitions for convenience of the reader .
we also prove some auxiliary lemmata that will be used in sections 3 and 4 , where we prove theorems [ th1 ] and [ th2 ] .
we denote by @xmath37 the set of all unit vectors in the euclidean space @xmath36 , and by @xmath38@xmath39@xmath40 an open hemisphere in the direction @xmath41 ; @xmath42 stands for the origin of @xmath36 . for any unit vector @xmath43
we let @xmath44 to be the orthogonal complement of @xmath45 in @xmath46 , i.e. the set of all @xmath47 such that @xmath48 ; here @xmath49 stands for a usual scalar product of @xmath50 and @xmath51 .
the notation for the orthogonal group @xmath26 and the special orthogonal group @xmath27 is standard .
the notation @xmath52 will be used for an orthogonal transformation acting in @xmath44 .
we agree to denote by @xmath53 the closed interval connecting @xmath50 and @xmath54 , i.e. all points of the form @xmath55 $ ] ; the shortest arc of the unit circle joining the points @xmath50 and @xmath54 on @xmath56 will be denoted by @xmath57 $ ] ; @xmath58 stands for the @xmath59-dimensional subspace that is a linear span of the linearly independent vectors @xmath60 ; @xmath61 is the linear subspace of smallest dimension containing a set @xmath62 .
also , for any set @xmath63 the notation @xmath64 is used for the projection of @xmath65 onto @xmath44 .
the _ shadow boundary _ of a convex polytope @xmath1 in direction @xmath45 will be denoted by @xmath66 , i.e. , @xmath67 ; here @xmath68 stands for the boundary of the projection @xmath69 .
given a set @xmath70 , @xmath71 denotes the smallets convex set containing @xmath70 .
a set @xmath72 is convex if for any two points @xmath50 and @xmath54 in @xmath65 the closed line segment @xmath53 joining them is in @xmath65 . a convex _ body _
@xmath73 is a compact convex set with a non - empty interior ( with respect to @xmath3 ) .
a _ convex polytope _
@xmath74 is a convex body that is a convex hull of finitely many points ( called the vertices of @xmath1 ) .
fix @xmath75 .
we say that a convex set @xmath70 is a _
@xmath6-dimensional convex polytope _ if there exists an affine @xmath6-dimensional subspace @xmath76 of @xmath3 such that @xmath77 and @xmath70 is a convex polytope relative to @xmath76 ( i.e. , the interior of @xmath70 is taken with respect to @xmath76 ) .
we will say that a subset @xmath78 of an open hemisphere @xmath38 , @xmath79 , is _ geodesically convex _ , if for any two points @xmath50 and @xmath54 in @xmath78 the arc of the unit circle @xmath57 $ ] joining them is in @xmath78 . we define a _ rigid motion _
@xmath80 acting in @xmath44 , @xmath81 , as a composition of an orthogonal transformation @xmath82 and a translation by @xmath83 , @xmath84 for all @xmath85 .
the supporting hyperplane @xmath86 to a convex body @xmath13 is defined as the hyperplane having common points with @xmath13 and such that @xmath13 lies in the closed half - space with the boundary @xmath86 .
the _ support function _ of a convex body is defined as @xmath87 for all @xmath88 .
if @xmath14 is a convex body @xmath14 containing the origin in the interior , its _ radial function _ is defined as @xmath89 for all @xmath88 . since @xmath90 and @xmath91 are homogeneous functions of degrees @xmath92 and @xmath93 correspondingly , it is enough to consider their values for @xmath94 , where both functions are continuous .
our first axillary result will be used in the proof of theorem [ th1 ] . [ length ] consider four points @xmath95 , such that for an open set @xmath96 of directions @xmath45 , @xmath97 .
then the intervals @xmath98 and @xmath99 are parallel and have equal length .
since any parallel translation preserves the length of the projections of the intervals , we may assume that @xmath100 .
let @xmath101 and @xmath102 be the unit vectors of directions of @xmath98 and @xmath99 , and @xmath103 and @xmath104 be their lengths . in this case
, the projections of @xmath70 and @xmath78 onto @xmath44 can be found as @xmath105 our goal is to prove that @xmath106 and @xmath107 or @xmath108 .
the condition on the equality of the lengths of projections can be written as @xmath109 or @xmath110 we claim that the right - hand side of the above identity is not an identically constant function of @xmath45 on the spherical cap @xmath111 , unless @xmath112 or @xmath113 . indeed ,
if the claim is false , then dividing the above identity by the lengths of vectors @xmath114 , @xmath115 , we see that the function @xmath116 is identically constant ( equal to @xmath117 ) on @xmath118 . for @xmath119
consider the level sets @xmath120 of functions @xmath121 and @xmath122 .
it is clear that @xmath123 , @xmath124 are the corresponding @xmath125-dimensional sub - spheres of @xmath56 . since @xmath126 , @xmath127 @xmath128 , @xmath129 , and @xmath130 is ( at most ) an @xmath131-dimensional subsphere of @xmath56 .
taking @xmath132 such that @xmath133 , and changing @xmath45 in @xmath134 , we see that @xmath135 remains constant while @xmath136 does not .
thus , @xmath137 is not constant on @xmath118 , and we obtain a contradiction . the claim is proven , and the proof of the lemma is finished .
the next statement shows that one may disregard the set of all directions for which at least two facets of the projections are orthogonal .
[ perp ] let @xmath138 and @xmath139 be two non - parallel @xmath125-dimensional subspaces in @xmath36 , such that @xmath140 .
then the set of directions @xmath141 , for which @xmath142 is orthogonal to @xmath143 , is a nowhere dense subset of @xmath56 .
assume that @xmath144 and @xmath145 , @xmath146 , where @xmath101 , @xmath147 , and @xmath148 are linearly independent vectors in @xmath3 , @xmath149 , @xmath150 .
by condition of the lemma , we have @xmath151@xmath152 , @xmath153 , @xmath154 , @xmath155 . here
@xmath156 , @xmath157 , @xmath158 , @xmath159 , and @xmath160 , @xmath161 .
let @xmath162 and @xmath163 be the normal vectors to @xmath142 and @xmath164 in @xmath44 . if @xmath165 , then @xmath166 , and , in the three - dimensional case , the condition of orthogonality @xmath167 can be written as @xmath168 to write out the condition in the case @xmath169 , we consider a linear transformation @xmath170 , such that @xmath171 . for all @xmath94 , @xmath172
, we choose @xmath173 with the matrix @xmath174 ( if we treat the @xmath175-th row @xmath176 of @xmath173 as an @xmath177-dimensional vector , then @xmath178 , @xmath179@xmath180@xmath92 , @xmath158 , @xmath177 ) .
if @xmath181 we put @xmath182 .
consider now the vectors @xmath183 .
the condition of the normals being orthogonal is @xmath184 where the normal vectors @xmath185 can be found as the generalized vector products , @xmath186 @xmath187 here @xmath188 stand for the @xmath189-th coordinate of the vectors @xmath190 and @xmath191 , @xmath192 is the @xmath189-th coordinate of the vector @xmath193 , and @xmath194 , @xmath195 , @xmath158 , @xmath196 , are the vectors of the standard basis in @xmath197 .
notice that the @xmath189-th coordinate of the normal @xmath162 is a rational function @xmath198 of the coordinates of @xmath45 , where @xmath199 is a polynomial in coordinates of @xmath45 .
similarly for @xmath163 , i.e. @xmath200 since we can choose the standard basis in @xmath3 in such a way that @xmath201 , we can assume that @xmath202 . multiplying each vector @xmath162 , @xmath203 by @xmath204 , we see that our condition on directions @xmath205 for which @xmath206 can be re - written as @xmath207 in other words , the desired set is @xmath208 , where @xmath209 is the set of zeros of @xmath210 on @xmath56 . by geometric considerations , @xmath210 is not identically zero ( it is enough to look at the directions @xmath45 that are close to the ones that are parallel to @xmath211 @xmath212 ) .
if the closure of @xmath213 has a non - empty interior then @xmath209 has a non - empty interior , i.e. , there exists a spherical cap @xmath214 of radius @xmath215 centered at @xmath216 . but this is impossible , for replacing @xmath217 with @xmath218 in the analytic expression for @xmath210 , we obtain an analytic function of variables @xmath219 in an open disc @xmath220 centered at the origin .
this contradicts the fact that the set of zeros of an analytic function of several real variables is of lebesgue measure zero ( cf .
@xcite ) .
thus , the closure of @xmath213 has an empty interior , which means that @xmath213 is nowhere dense .
the lemma is proved
. the following elementary result will be crucial in the proof of theorem [ th2 ] .
[ lines ] let @xmath118 be an open subset of @xmath56 and let @xmath221 be four lines in @xmath222 , not passing through the origin , such that for any @xmath223 the subspace @xmath44 intersects each line at a single point @xmath224 .
assume also that @xmath225 for any @xmath223 .
\1 ) if @xmath226 and @xmath227 are parallel , then all four lines are parallel . in addition
, there exists a translation @xmath147 , such that @xmath228 and @xmath229 or @xmath230 and @xmath231 .
\2 ) if @xmath226 and @xmath227 are not parallel , then one of the following holds : @xmath232 or @xmath233 .
\3 ) if @xmath226 and @xmath227 are not parallel and @xmath234 , then one of the following holds : @xmath235 , or @xmath236 , or @xmath237 , or @xmath238 .
we start the proof with some elementary algebraic observations .
parameterize each line @xmath239 , such that @xmath240 , @xmath241 and @xmath242 .
notice , that the choice of the directional vectors @xmath243 is determined up to a sign , and the value of the parameter @xmath103 for the points of intersection can be found from the condition @xmath244 , i.e. @xmath245 .
hence , the condition of the lemma can be re - written as @xmath246 in other words , @xmath247 where @xmath248 @xmath249 @xmath250 and , due to the fact that @xmath251 , the fractions above are not reducible . hence , @xmath252 and @xmath253 for all @xmath254 . since all polynomials are homogeneous of degree @xmath255 , we can extend @xmath256 , @xmath257 , @xmath258 , @xmath259 to the whole space @xmath3 and , by analiticity , we have @xmath252 and @xmath253 for all @xmath88 .
we prove 1 ) .
assume that @xmath226 and @xmath227 are parallel , i.e. , that @xmath260 .
then @xmath261 reads @xmath262 .
the left - hand side of this identity is divisible by @xmath263 , hence the right - hand side must be divisible as well .
this gives @xmath264 , i.e. , all four lines are parallel .
now we will use @xmath265 .
it reads as @xmath266 or , using the fact that @xmath267 , @xmath268 we claim that @xmath269 . if @xmath270 , dividing all terms in the above expression by @xmath271 , we see that @xmath272 must be divisible by @xmath273 . in other words , @xmath274 for some @xmath275 . since all @xmath276
, we have @xmath277 , and the claim is proven .
we see now that ( [ algebra ] ) has reduced to @xmath278 or to @xmath279 the first part of the lemma is proven .
we prove 2 ) .
we will consider the first four possibilities , other four options are obtained by changing indices @xmath280 .
comparing the denominators in ( [ fraction ] ) , we see that @xmath281 , @xmath282 . writing out the numerators in ( [ fraction ] ) , we have @xmath283 @xmath284 @xmath285 @xmath286 @xmath287 since the second term must be divisible by @xmath288 and the third one must be divisible by @xmath289 , we see that @xmath290 or @xmath291 for some @xmath292 , @xmath293 ( observe that both conditions may not hold simulteneously , for , otherwise , summing them up , we would obtain that @xmath294 is parallel to @xmath295 ) and @xmath296 or @xmath297 for some @xmath298 , @xmath299 ( again , both conditions may not hold simulteneously , otherwise @xmath300 would be parallel to @xmath301 ) . if @xmath290 , then @xmath302 , for , the parametric equation of @xmath303 becomes @xmath304 , @xmath305 , which is the same as the one of @xmath306 , @xmath307 , after taking @xmath308 . arguing similarly , we see that @xmath309 , @xmath310 .
this finishes the proof of 2 ) . in order to prove 3 ) , it remains to exclude two cases @xmath311 , @xmath312 , and @xmath313 , @xmath314 , provided @xmath315 ( the cases obtained by changing the indices @xmath280 are excluded similarly ) .
we will consider the first case and the exclusion of the second case is similar . to this end , assume that @xmath316 , @xmath317 , and @xmath318 , @xmath319 .
then , condition ( [ main ] ) reads as @xmath320 it is satisfied only , provided @xmath321 ( observe that ( [ oshibka ] ) holds in the case @xmath169 , @xmath322 , and @xmath323 ) .
we show that ( [ oshibka ] ) leads to a contradiction in the case @xmath315 .
we may assume that @xmath324 .
we observe that @xmath325 .
passing to the common denominator in ( [ oshibka ] ) and using analiticity , we have @xmath326 @xmath327 we can assume that one of the numbers @xmath328 , @xmath329 , @xmath330 , @xmath331 , is not zero ; without loss of generality , @xmath332 .
then , the expression @xmath333 is divisible by @xmath288 , and , as a consequence , @xmath334 is divisible by @xmath288 .
this gives @xmath335 for some @xmath336 .
similarly , since @xmath337 is divisible by @xmath289 , the expression @xmath338 is divisible by @xmath289 as well , and we have @xmath339 for some @xmath340 . combining ( [ raz ] ) and ( [ dva ] ) together , and using the fact that @xmath294 is not parallel to @xmath300 , we obtain @xmath325
. finally , denote @xmath341 we know that for every @xmath342 these vectors belong to the same two - dimensional subspace of @xmath343 . using ( [ oshibka ] ) and the analyticity of @xmath344 and @xmath345 , we see that @xmath346 and @xmath347 for all @xmath348 , i.e. , @xmath344 and @xmath345 belong to @xmath44 .
moreover , due to ( [ oshibka ] ) they are orthogonal to each other . by taking @xmath349 ,
we obtain two vectors that are orthogonal and parallel to each other at the same time .
hence , at least one of them is zero , and we obtain a contradiction .
the lemma is proven .
let @xmath1 be a polytope in @xmath350 . to formulate our next lemma
, we denote by @xmath351 the open set of directions @xmath45 such that @xmath44 is not passing through vertices of @xmath1 .
[ dual ] let @xmath138 and @xmath139 be two @xmath352-dimensional subspaces of @xmath3 parallel to the facets of a polytope @xmath1 .
then the set of directions @xmath223 such that @xmath353 and @xmath354 are orthogonal , is a nowhere dense subset of @xmath56 .
we can assume that @xmath355 .
consider the dual polytope @xmath356 .
it is well - known that @xmath357 ( see , for instance , @xcite , formula ( 0.38 ) , p. 22 ) . by condition of the lemma , @xmath254 is not parallel to any facet of @xmath358 . hence , the pre - image of any vertex of @xmath359 under the projection is a vertex of @xmath356 . on the other hand , under the duality the facets of @xmath1
become the vertices of @xmath358 and the radius - vectors of the vertices of @xmath356 are parallel to the normal vectors @xmath360 , @xmath361 of @xmath362 and @xmath363 .
hence , the vectors @xmath364 and @xmath365 are parallel to the vertices of @xmath359 , and the condition of the orthogonality of facets of the sections of @xmath366 is equivalent to @xmath367 .
this reads as @xmath368 the last equation is a polynomial function of coordinates of @xmath45 . repeating the argument which is similar to the one from lemma [ perp ]
, we see that the set of directions satisfying the above condition is nowhere dense on @xmath56 .
the lemma is proven .
we will use the well - known minkowski theorem ( see , for example , @xcite , theorem 9 , p. 282 ) .
[ minkowski ] suppose that @xmath13 and @xmath14 are polytopes in @xmath369 , and suppose also that the facet unit normals @xmath370 and corresponding facet areas @xmath371 coincide , then @xmath13 coincides with @xmath14 up to a translation .
our last auxiliary result in this section is [ hedgehog ] let @xmath372 and let @xmath210 and @xmath373 be two continuous real - valued functions on @xmath56 .
assume that for any @xmath189-dimensional subspace @xmath138 and some vector @xmath374 , the restrictions of @xmath210 and @xmath373 onto @xmath375 satisfy @xmath376 @xmath377 or @xmath378 @xmath377 .
then there exists @xmath379 such that @xmath380 @xmath381 or @xmath382 @xmath381 .
we remark also that if all @xmath383 are zero , then @xmath384 ( see also lemma 1 from @xcite , page 3431 ) .
we start with the case @xmath385 . we will show that for all directions @xmath94 , the projections of polytopes onto @xmath44 coincide up to a translation and a reflection in the origin .
this will be achieved in three steps that can be briefly sketched as follows .
_ step 1_. let @xmath386 be a closed set of directions @xmath45 that are parallel to facets of @xmath1 or @xmath2 , and let @xmath387 be any connected component of @xmath388 .
we will prove that , given any two vectors @xmath389 and @xmath390 in @xmath387 , we have @xmath391 and @xmath392 ; see lemma [ shadow ] .
_ step 2_. we will prove that the edges of the shadow boundaries @xmath66 and @xmath393 , @xmath223 , are in a bijective correspondence .
we will apply lemma [ length ] to all corresponding edges of @xmath66 and @xmath393 to conclude that they are parallel and have equal length .
_ step 3_. we will show that the corresponding facets of projections @xmath69 and @xmath394 , @xmath254 , are pairwise parallel .
we will apply minkowski s theorem to conclude that @xmath69 and @xmath395 , @xmath254 , coincide up to a translation and a reflection in the origin .
then , we will use the density argument to conclude that the last statement holds for all directions @xmath94
. _ step 1_. fix any facet @xmath396 of @xmath1 or @xmath2 with an outer unit normal @xmath397 .
any direction @xmath43 which is parallel to @xmath396 belongs to @xmath398 .
since polytopes have a finite number of facets , the set @xmath399 of all such directions @xmath45 is a union of a finite number of unit @xmath125-dimensional subspheres of @xmath56 .
denote @xmath400 .
since @xmath399 is closed , @xmath118 is open .
fix any @xmath223 .
any vertex @xmath401 is a projection of some vertex ( not an edge ) , otherwise @xmath45 is parallel to some facet .
the same holds for edges of @xmath69 , i.e. a pre - image of any edge of @xmath69 is an edge of @xmath1 , otherwise @xmath45 is parallel to some @xmath402-dimensional face of @xmath1 .
let @xmath387 be any non - empty linearly connected component of @xmath118 .
observe that it is contained in an open hemisphere @xmath403 for some @xmath41 .
note also that @xmath387 is geodesically convex , i.e. , for any two directions @xmath389 and @xmath390 in @xmath387 there exists an arc @xmath404 $ ] of a unit circle of @xmath56 such that @xmath404\subset u_1 $ ] . indeed ,
since the boundary @xmath405 of @xmath387 is a finite union of closed pieces of @xmath125-dimensional unit subspheres of @xmath56 , @xmath406 , we see that @xmath407 , where @xmath408 is a finite convex intersection of @xmath189 half - spaces defined by hyperplanes , passing through the origin and containing @xmath409 . since @xmath408 is convex , the interval @xmath410 connecting @xmath389 and @xmath390 belongs to @xmath408 .
the projection of this interval onto @xmath56 is an arc @xmath404 $ ] of the unit circle , which is contained in @xmath387 .
assume the opposite , there exist two distinct vectors @xmath411 , such that @xmath413 .
then there exists @xmath414 , but @xmath415 , and we may construct two half - lines @xmath416 and @xmath417 ( or @xmath418 , if necessary ) to obtain @xmath419 take a small enough @xmath420 , such that the ball @xmath421 intersects only faces of @xmath1 that contain @xmath50 ( see figure 1 ) . choose @xmath422 and @xmath423 .
notice that @xmath424 , otherwise @xmath425 , which is not possible since @xmath389 is parallel to @xmath426 .
we project the interval @xmath427\}$ ] onto @xmath56 and obtain an arc @xmath428 of the unit circle @xmath429 . on the other hand
, there exists a point @xmath430 , such that @xmath431 for some @xmath432 , where @xmath433 is a facet of @xmath1 containing @xmath430 .
we have @xmath434 , which contradicts the fact that @xmath387 is geodesically convex .
_ step 2_. our next goal is to show that the edges of shadow boundaries @xmath66 and @xmath393 , @xmath435 , are in a bijective correspondence .
moreover , we will prove that the corresponding edges are parallel and have equal length .
assume that for a fixed @xmath436 the projection @xmath69 has @xmath6 vertices @xmath437 , @xmath438 which are the projections of @xmath6 vertices @xmath439 of @xmath1 . since @xmath440 , every vertex @xmath441 is mapped onto some vertex @xmath442 and every edge of @xmath69 is mapped onto some edge of @xmath395 ( notice also that the pre - image of any vertex of @xmath443 is a vertex of @xmath2 and the pre - image of any edge of @xmath395 is an edge of @xmath2 ) .
this implies that for any @xmath435 we obtain a bijective correspondence @xmath444 between the set of all vertices @xmath445 of the shadow boundary @xmath66 and the set of all vertices @xmath446 of the shadow boundary @xmath393 .
take a closed spherical cap with a non - empty interior @xmath447 .
for any @xmath448 we have a map @xmath449 , such that @xmath450 , and @xmath451 is a permutation of the set @xmath452 satisfying @xmath453 .
the set of all such possible maps @xmath444 is finite .
we have @xmath454 where @xmath455 is the set of all permutations of @xmath452 .
observe that each @xmath456 is a closed set ( it might be empty ) .
indeed , let @xmath457 be a convergent sequence of points of a non - empty @xmath456 , and let @xmath458 .
we have @xmath459 , i.e. , @xmath460 where @xmath461 is independent of @xmath462 .
for each @xmath463 extend the operator @xmath80 acting from @xmath44 to @xmath44 to the operator @xmath464 acting from @xmath3 to @xmath3 as @xmath465 . here
@xmath466 is defined as @xmath467 , @xmath468 , where @xmath469 .
writing ( [ close ] ) in terms of @xmath464 , @xmath470 and passing to the limit as @xmath471 ( by considering the corresponding subsequences if necessary ) , we obtain @xmath472 this means that @xmath473 and @xmath456 is a closed . by the baire category theorem ( see , for example , @xcite , pages 42 , 43 ) there exists a permutation @xmath474 such that the interior @xmath475 of @xmath476 is non - empty .
( this could be also easily seen as follows .
enumerate @xmath477 , and take the first set @xmath478 .
if its interior is empty , it is nowhere dense , and for every open spherical cap @xmath479 in @xmath480 there exists a smaller cap @xmath481 that is free of points of @xmath478 , @xmath482 .
repeat the procedure with @xmath483 and @xmath481 instead of @xmath478 and @xmath479 .
after finitely many steps , unless we meet some @xmath456 with a non - empty interior , we will optain a spherical cap that does not intersect @xmath480 , which is impossible ) .
observe that if @xmath484 and @xmath485 are connected by an edge @xmath486 , such that @xmath487 is an edge of @xmath69 , @xmath488 , then @xmath489 is an edge of @xmath2 .
now , we can apply lemma [ length ] to all such pairs of edges @xmath490 and @xmath491 for any @xmath492 .
we see that these edges are parallel and have equal length .
thus , all corresponding edges belonging to the shadow boundaries @xmath66 and @xmath393 , @xmath493 , are parallel and have equal lengths . applying lemma [ shadow ] , we conclude that the last statements holds for @xmath387 instead of @xmath475 .
_ step 3_. we will show that the projections of both polytopes in the directions of @xmath387 coincide up to a translation and a reflection in the origin .
to do this , we will use minkowski s theorem about uniqueness ( up to a translation ) of polytopes with parallel facets having the same volume .
our polytopes will be @xmath69 and @xmath395 , @xmath435 .
fix any @xmath436
. then @xmath69 and @xmath395 are two @xmath352-dimensional polytopes in @xmath44 , such that @xmath494 .
since the map @xmath80 is a bijection between the sets of all facets of @xmath69 and @xmath395 , for any facet @xmath138 of @xmath69 there exists a unique facet @xmath495 of @xmath443 such that @xmath496 .
we will show at first that @xmath138 is parallel to @xmath495 .
consider the @xmath125-dimensional affine subspaces @xmath497 , @xmath498 of @xmath44 such that @xmath499 and @xmath500 .
we claim that @xmath498 is parallel to @xmath501 .
in other words , the outer unit normals @xmath360 , @xmath502 of @xmath497 and @xmath498 satisfy @xmath503 .
to prove the claim , observe that there exist @xmath504 linearly independent non - zero directional vectors @xmath505 of edges of @xmath138 having the common vertex , such that @xmath506 is parallel to @xmath497 ( here , a directional vector of an edge is any non - zero vector parallel to the edge ) .
the directional vectors of edges @xmath505 are the projections of directional vectors of the edges @xmath507 .
the directional vectors @xmath508 of the corresponding edges are parallel , i.e. @xmath243 is parallel to @xmath509 for @xmath510 .
the same holds true for their projections onto @xmath44 , @xmath511 is parallel to @xmath512 for @xmath510 .
we conclude that for the @xmath125-dimensional affine subspace @xmath498 containing @xmath495 we have @xmath513 and @xmath498 is parallel to @xmath514 .
the claim is proved . since @xmath80 is an isometry , we have @xmath515 .
now assume that both polytopes @xmath69 and @xmath395 have @xmath516 facets @xmath517 and @xmath518 with corresponding outer normals @xmath519 and @xmath520 .
we will show that up to a nowehere dense subset of directions @xmath45 ( defined below ) the projections @xmath69 and @xmath395 are translates of each other ( up to a reflection in the origin ) . to be able to apply minkowski s theorem to @xmath69 and @xmath395 ( or @xmath521 and @xmath395 ) we have to show that @xmath522 for all @xmath523 , or @xmath524 for all @xmath523 .
assume that the last statement is not true , and let @xmath525 and @xmath526 , be two pairs of facets such that @xmath527 , @xmath528 , @xmath529 , but @xmath530 , ( since @xmath80 is an isometry , @xmath531 ) .
consider two cases , @xmath532 and @xmath533 .
the first case @xmath532 is impossible , for @xmath534 yields @xmath533 . here
@xmath535 stands for the transpose operator of @xmath82 , and we used the fact that @xmath536 due to @xmath537 . to exclude the case @xmath533 , we recall that the pre - image of any facet of @xmath69 or @xmath395 is an @xmath125-dimensional face of @xmath1 or @xmath2 respectively .
we apply lemma [ perp ] to the pairs of subspaces that are parallel to all pairs of @xmath125-dimensional faces of @xmath1 and @xmath2 .
we obtain a closed nowhere dense subset @xmath538 , such that the normals of the corresponding facets of @xmath69 and @xmath395 , @xmath539 , are orthogonal .
now we may repeat our considerations for the case @xmath532 on @xmath540 instead of @xmath387 .
we obtain that for every @xmath541 only one of the choices @xmath522 for all @xmath523 , or @xmath524 for all @xmath523 , holds .
we conclude that for all @xmath541 , we have @xmath522 for all @xmath523 , or @xmath524 for all @xmath523 .
also , recall that @xmath542 for all @xmath523 , where @xmath543 , @xmath544 ; now we may apply theorem [ minkowski ] to @xmath69 and @xmath395 or to @xmath521 and @xmath395 .
thus , for every @xmath541 there exists a vector @xmath545 , such that @xmath546 or @xmath547 . we can repeat the above argument for every connected component of @xmath118 .
we see that for every @xmath548 the supporting functions @xmath549 and @xmath550 satisfy @xmath551 or @xmath552 here we used the fact that @xmath553 @xmath554 , see ( @xcite , ( 0.21 ) , p.17 ) .
it is not difficult to see that one of the above equalities ( or both ) hold for every @xmath94 . indeed , since @xmath555 is nowhere dense on @xmath56 , for any open neighborhood @xmath556 of any @xmath557 we have @xmath558 . hence , there exists a sequence @xmath559 , such that @xmath560 . by taking a convergent subsequence if necessary ,
we conclude that ( the argument is very similar to the one in the proof of lemma 6 , p. 3435 ,
@xcite ) there exists a limit @xmath561 for @xmath557 , such that ( [ plus ] ) or ( and ) ( [ minus ] ) holds .
to finish the proof in the case @xmath562 , we apply theorem [ hedgehog ] with @xmath563 , and @xmath564 , @xmath565 .
we obtain that there exists @xmath31 such that @xmath566 for all @xmath567 , or @xmath568 for all @xmath569 . using the well - known properties of the support functions ( @xcite , pp .
16 - 18 ) we obtain the desired result in the case @xmath562 .
we use induction on @xmath6 .
let @xmath7 be any @xmath571-dimensional subspace of @xmath3 .
we apply the result for @xmath572 to the bodies @xmath4 and @xmath5 and their projections @xmath573 and @xmath574 for all @xmath575-dimensional subspaces @xmath576 .
we obtain that @xmath577 or @xmath578 for all @xmath7 .
we proceed and the result follows after finitely many steps .
the proof of theorem [ th1 ] is completed .
we will show that for all directions @xmath94 , the sections of polytopes by @xmath44 coincide up to a reflection in the origin .
this will be achieved in four steps , which can be briefly described as follows .
_ step 1_. let @xmath386 be a closed set of directions @xmath45 for which @xmath44 contains any of the vertices of @xmath1 or @xmath2 , and let @xmath387 be any connected component of @xmath388 .
we will prove that given any two directions @xmath389 and @xmath390 in @xmath387 , the subspaces @xmath579 and @xmath580 intersect the same set of edges of @xmath1 ( and @xmath2 ) ; see lemma [ edges ] .
denote by @xmath581 and @xmath582 the sets of edges of @xmath1 and @xmath2 that are intersected by @xmath44 , @xmath435 .
denote also by @xmath583 and @xmath584 the sets of lines containing the intersected edges from @xmath585 and @xmath586 .
_ step 2_. we will prove that there exists an open non - empty subset @xmath475 of @xmath387 such that for all directions @xmath488 all rigid motions @xmath80 look alike " .
more precisely , for all @xmath488 and for any @xmath587 the vertices @xmath588 of sections @xmath366 are all mapped into the vertices of @xmath589 of the form @xmath590 , where the line @xmath591 is fixed and @xmath592 is independent of @xmath45 . _ step 4_. we will apply minkowski s theorem to the facets of the sections @xmath594 and @xmath595 and will prove that for any @xmath94 these sections coincide up to translation and a reflection in the origin .
next , using _
step 3 _ we will show that the sections coincide up to a reflection in the origin .
_ step 1_. consider any vertex @xmath598 of @xmath1 or @xmath2 . if a hyperplane @xmath44 contains @xmath598 then @xmath599 .
hence , the set of all such directions @xmath45 is a sub - sphere @xmath600 .
since both polytopes have a finite number of vertices , the union @xmath399 of all such sub - spheres is closed , hence its complement @xmath601 is open in @xmath56 .
assume that @xmath606 ( see figure 2 ) .
then , there exists an edge @xmath53 such that @xmath607 , @xmath608 , for some point @xmath609 .
assume also that @xmath610 , where @xmath516 is a line @xmath611 , @xmath612 , @xmath613 .
to prove the claim consider a continuous path @xmath614 along the shortest arc @xmath615$]@xmath616 , such that @xmath617 , @xmath618 ( since @xmath387 is geodesically convex , @xmath615\in u_1 $ ] ) . for any @xmath619
the intersection of @xmath614 with @xmath516 can be found as @xmath620 , such that @xmath621 and @xmath622 ( if @xmath623 take @xmath624 and consider the point @xmath625 for some @xmath626 small enough ) . since @xmath627 is continuous , there exists @xmath628 , such that @xmath629 ( or @xmath54 ) .
this would imply that @xmath630 , which leads to a contradiction .
_ step 2_. for any @xmath436 the vertices @xmath484 of @xmath366 are the intersections of @xmath44 with lines @xmath587 and the vertices @xmath632 of @xmath595 are the intersections of @xmath44 with lines @xmath633 . since the rigid motion @xmath80 maps the vertices of @xmath366 onto the vertices of @xmath595 , there exists a one to one correspondence between these sets of vertices .
hence , there also exists a one to one correspondence between the lines in @xmath14 and @xmath593 .
let @xmath634@xmath635 and @xmath636@xmath637 .
then for every @xmath435 , the rigid motion @xmath80 maps any vertex @xmath588 of @xmath594 into the vertex @xmath638 of @xmath589 , @xmath639 .
hence , there is a permutation @xmath640 of the set @xmath641 and the map @xmath642 such that @xmath643 , @xmath644 .
observe each @xmath456 is a closed set ( it might be empty ) . indeed , let @xmath457 be a convergent sequence of points of a non - empty @xmath456 , and let @xmath458 .
then @xmath648 .
in other words , for the corresponding vertices @xmath649 and @xmath650 we have @xmath651 where @xmath461 is independent of @xmath462 . arguing as in the case of projections ( see _ step 2 ) _ in the proof of theorem [ th1 ] ) , we extend the rigid motion @xmath652@xmath653 to the one acting on the whole space @xmath654@xmath655 , @xmath463 , by the formula @xmath465 . here
@xmath466 is defined as @xmath467 , @xmath468 , where @xmath469 . writing ( [ vertline ] ) in terms of @xmath464 we have @xmath656 now we pass to the limit in the last equality as @xmath471 ( taking the subsequences if necessary ) , and using the fact that the norms of @xmath657 are uniformly bounded , we obtain @xmath658 hence , @xmath659 and @xmath473 .
this means that @xmath456 is a closed . by the baire category theorem ( we argue as in _ step 2 ) _ in the proof of theorem [ th1 ] ) there exists a permutation @xmath474 such that the interior @xmath660 is non - empty .
hence , ( [ perm ] ) holds with @xmath661 and @xmath662 .
_ step 3_. by lemma [ edges ] the set of edges intersected by @xmath44 , @xmath488 , coincides with the set of edges intersected by @xmath44 , @xmath435 . to show that the corresponding lines in @xmath14 and @xmath593 coincide up to a reflection in the origin
, we will consider two cases : at least two lines in @xmath14 are not parallel , or all lines in @xmath14 are parallel .
let @xmath488 and let @xmath461 be as in ( [ perm ] ) .
if at least two lines from @xmath14 , say @xmath663 , are not parallel , we apply the second part of lemma [ lines ] to the pair of lines @xmath664 with vertices @xmath665 , @xmath666 and the corresponding pair of lines @xmath667@xmath668 with vertices @xmath669 , @xmath670 .
we see that the lines coincide up to a reflection in the origin .
considering all corresponding pairs of pairwise non - parallel lines @xmath671 and @xmath672 ( or @xmath673 and @xmath674 ) , @xmath675 , @xmath255 , @xmath676 , with the corresponding vertices , we conclude that the lines from @xmath14 and @xmath593 coincide up to a reflection in the origin .
in other words , for all @xmath195 , @xmath677 , @xmath678 .
suppose now that for all @xmath493 , @xmath44 intersects only parallel lines in @xmath14 .
first , we will show that there exists @xmath31 such that @xmath679 second , we will prove that @xmath384 or @xmath680 for all @xmath681 . to prove ( [ shift ] ) we let @xmath682 and @xmath683 be any two parallel lines from @xmath14 .
applying the first part of lemma [ lines ] to @xmath684 , @xmath685@xmath686 , @xmath687 , @xmath688@xmath668 , and the corresponding vertices , we see that the lines @xmath689 are parallel and @xmath690 , for some @xmath691 ( up to a change of indices @xmath692 ) .
consider now a triplet of lines @xmath682 , @xmath693 , and @xmath683 , such that two two - dimensional faces of @xmath1 have a common edge contained in @xmath682 and they have other parallel edges contained in @xmath693 and @xmath683 .
let @xmath488 .
if @xmath484 are the vertices of @xmath594 , then the vertices @xmath694 , @xmath695 , and @xmath696 of @xmath589 should belong to the edges of @xmath2 .
moreover , due to the congruence of @xmath594 and @xmath589 the edge containing @xmath697 should be common to the corresponidng two - dimensional faces of @xmath2 .
we apply the first part of lemma [ lines ] to the pairs @xmath698 and @xmath699 ) and to the corresponding vertices .
observe that the line @xmath682 must be translated by the same vector in both pairs @xmath700 and @xmath698 , otherwise the pair of lines @xmath701 , @xmath688 is not a translation of the pair @xmath702 , which contradicts the first part of lemma [ lines ] .
hence , @xmath703 , where @xmath147 is the same as for the pairs @xmath700 , @xmath687 , @xmath688 . taking the corresponing triplets of lines in @xmath14 and @xmath593 after finitely many steps we have ( [ shift ] ) .
let @xmath704 be any line among the parallel lines in @xmath14 .
consider all edges of @xmath1 that have a common vertex @xmath705 with @xmath706 , say , @xmath707 , @xmath708 .
consider also all lines @xmath709 containing @xmath707 , and let @xmath710 be the corresponding connected component of @xmath118 for which @xmath44 intersect @xmath682 and @xmath709 , for all @xmath711 .
observe that @xmath712 , @xmath713 .
we apply the third part of lemma [ lines ] to the pairs @xmath714 and @xmath715 and the corresponding vertices @xmath716 , @xmath717 and @xmath718 , @xmath719 . here
the lines @xmath682 and @xmath720 belong to the corresponding sets of lines @xmath721 and @xmath722 and @xmath461 is the fixed permutation analogous to the one in ( [ perm ] ) . if lemma [ lines ] yields @xmath723 , @xmath724 , for all @xmath708 , we claim that @xmath384 . indeed , in this case both polytopes @xmath1 and
@xmath2 have a common solid angle @xmath725 containing @xmath705 .
let @xmath726 , @xmath727 , be a supporting hyperplane to both @xmath1 and @xmath2 with an inner ( with respect to @xmath1 ) unit normal @xmath728 .
we will consider two cases , @xmath729 and @xmath730 , and show that both are impossible ( by changing @xmath7 slightly we can assume that @xmath731 ) . by ( [ shift ] ) , @xmath732 and @xmath733 . if @xmath729 , then @xmath2 should lie in the translated half - space @xmath734 , where @xmath408 is the half - space with boundary @xmath7 containing @xmath1 and @xmath2 .
since @xmath2 contains @xmath735 this is impossible , unless @xmath384 . if @xmath730 , then @xmath2 contains @xmath725 and @xmath736 , which contradicts the convexity of @xmath2 .
hence , @xmath384 .
if for all @xmath737 there exists @xmath738 , @xmath708 , such that lemma [ lines ] yields @xmath680 , @xmath739 , we have @xmath680 for all @xmath681 .
thus , the corresponding lines from @xmath14 and @xmath593 coincide up to a reflection in the origin .
consider any section @xmath740 , @xmath436 , and let @xmath698 be a pair of lines in @xmath14 such that @xmath743 is an edge of @xmath740 .
since the edge @xmath743 is obtained as an intersection of @xmath44 with the two - dimensional face of @xmath1 , we have @xmath744 .
let @xmath461 be a permutation from ( [ perm ] ) . applying the first or the third part of lemma [ lines ] to the pairs of lines @xmath698 and @xmath699 ( depending on whether they are parallel or not ) and to the corresponding vertices @xmath224 , @xmath745 , @xmath746 ,
@xmath747 , @xmath488 , we know that @xmath748 or @xmath749 ( up to a change of indices @xmath750 ) . hence , for any @xmath435 , the edge @xmath751 of @xmath752 satisfy @xmath753 @xmath754 or @xmath753 @xmath755 . in either case ,
the vectors that are parallel to these edges are preserved up to a sign . now consider a facet @xmath138 of @xmath756 and the corresponding facet @xmath757 of @xmath752 .
since the pairs of the corresponding parallel edges contained in @xmath138 and @xmath495 coincide up to a reflection in the origin , the @xmath125-dimensional affine planes containing the corresponding facets @xmath138 and @xmath495 are parallel .
in addition , @xmath758@xmath759 . to be able to apply minkowski s theorem we need to show that all outer normals to the corresponding parallel facets of @xmath740 and @xmath741 coincide or all are opposite ( in the last case we will consider the facets of @xmath756 and @xmath760 ) . by lemma [ dual ]
the set @xmath761 of directions for which the outer normal vectors of at least two facets of sections are orthogonal , is nowhere dense on @xmath56 .
hence , arguing as in the case of projections we see that for all @xmath541 we may apply minkowski theorem to obtain that @xmath762 or @xmath763 to prove the claim , we notice that if all corresponding lines in @xmath14 and @xmath593 are reflections of each other , then @xmath769 ( in this case the corresponding vertices of the sections are reflections of each other ) . by _
step 3 _ the lines in @xmath14 and @xmath593 coincide up to a reflection in the origin , hence , _ all _ corresponding lines are reflections of each other , provided @xmath770
. we will show that if @xmath771 then all corresonding lines in @xmath14 and @xmath593 are reflections in the origin of each other as well .
let @xmath772 .
consider the sets @xmath773 and @xmath774 of all vertices of @xmath594 an @xmath589 , @xmath435 .
the set @xmath775 is the union of @xmath776 and @xmath777 , where @xmath778 is the set of vertices of @xmath594 belonging to @xmath779 and @xmath780 is the set of vertices of @xmath594 belonging to @xmath781 .
similarly , the set @xmath782 is the union of @xmath776 and @xmath783 , where and @xmath784 is the set of vertices of @xmath589 belonging to @xmath785 .
we have to show that @xmath786 .
we will prove at first that the set @xmath787 is invariant under the reflection in the origin .
let @xmath788 be any vertex .
then @xmath789 , @xmath790 , is the vertex of @xmath741 and @xmath740 . by ( [ refl ] ) , @xmath791 is the vertex of @xmath594 and the line @xmath792 is in @xmath781 ( if @xmath793 again , then @xmath794 is the vertex of @xmath594 and the segment joining @xmath598 and @xmath794 is in @xmath594 , hence @xmath791 is not the vertex ) . by
step 3 _ there must be a line @xmath795 and @xmath796 is the vertex of @xmath589 . by ( [ refl ] )
@xmath797 is also a vertex of @xmath594 .
similarly , by ( [ refl ] ) @xmath798 must be a vertex of @xmath589 and @xmath799 is a vertex of @xmath594 ( for , @xmath798 must belong to the line in @xmath785 ) , hence @xmath797 is the vertex of @xmath589 .
thus , @xmath797 is the vertex of both @xmath594 and @xmath589 and @xmath800 .
by _ step 3 _ , for every vertex @xmath801 there exists a vertex @xmath802 such that @xmath803 . since @xmath804@xmath805
, we obtain @xmath806@xmath807 , and @xmath786 .
we proved that if @xmath772 then @xmath808 .
this means that ( [ refl ] ) is possible only provided @xmath764@xmath765@xmath766 or @xmath767 . the claim is proved ( to show that ( [ refl1 ] ) is possible only provided @xmath768@xmath765@xmath756 or @xmath767 it is enough to consider @xmath1 and @xmath8 instead of @xmath1 and @xmath2 ) .
thus , @xmath809 the proof of _ step 4 _ is finished .
we have shown that for all @xmath810 , we have @xmath811 for all @xmath812 or @xmath813 for all @xmath812 .
since the radial functions are continuous we can prove that the same holds for any @xmath557 , and hence for all @xmath94 .
finally , we can apply theorem [ hedgehog ] with @xmath596 , @xmath597 , to conclude that @xmath811 for all @xmath814 or @xmath815 for all @xmath814 .
theorem [ th2 ] is proved in the case @xmath562 .
we use induction on @xmath6 .
let @xmath7 be any @xmath571-dimensional subspace of @xmath3 .
we apply the result for @xmath572 to the bodies @xmath816 and @xmath817 and their sections @xmath818 and @xmath819 for all @xmath575-dimensional subspaces @xmath576 .
we obtain that @xmath820 or @xmath821 for all @xmath7 .
we proceed and the result follows after finitely many steps .
the proof of theorem [ th2 ] is finished . , _ on symmetries of projections and sections of convex bodies _ , springer contributed volume discrete geometry and symmetry " dedicated to karoly bezdek and egon schulte on the occasion of their 60-th birthdays , to appear . | let @xmath0 and let @xmath1 and @xmath2 be two convex polytopes in @xmath3 .
assume that their projections , @xmath4 , @xmath5 , onto every @xmath6-dimensional subspace @xmath7 , are congruent . in this paper
we show that @xmath1 and @xmath2 or @xmath1 and @xmath8 are translates of each other .
we also prove an analogous result for sections by showing that @xmath9 or @xmath10 , provided the polytopes contain the origin in their interior and their sections , @xmath11 , @xmath12 , by every @xmath6-dimensional subspace @xmath7 , are congruent . |
the clustering of galaxies reflects both the spatial distribution of dark matter , which is dependent upon cosmological parameters , and the complex physics which governs the creation and evolution of galaxies within their host dark matter halos .
clustering measures can therefore constrain both cosmology ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ) and galaxy formation models , including parameterized models like the halo occupation distribution ( hod ) models ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) or more physical semi - analytic models and hydrodynamic simulations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
in particular , the detailed dependence of clustering on galaxy properties such as luminosity and color at different redshifts provides a wealth of information on galaxy evolution .
the formation and evolution of dark matter halos is straightforward to predict and has been characterized with large n - body simulations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , but the details of how baryons in galaxies populate their parent dark matter halos are less well understood and should depend on the detailed physics of gas accretion , star formation , and feedback processes .
comparisons of the observed dependence of clustering on galaxy properties at different epochs with theoretical models and simulations can illuminate the relative importance of various physical processes involved in galaxy evolution . on scales of @xmath4 the galaxy two - point correlation function
is roughly a power law , @xmath5 , and has long been known to depend on galaxy properties , with one of the strongest dependencies being galaxy type .
the division of the overall galaxy population into two distinct types has been seen clearly as a function of galaxy morphology , color , and spectral type , where galaxies tend to be either early / bulge - dominated / red / non - star forming or late / disk - dominated / blue / star forming ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
restframe color has been found to correlate well with both spectral type and morphology , with some scatter . in this paper
we focus on the dependence of clustering on restframe color , as this quantity can be determined more robustly for large samples of galaxies at both low and high redshift than , for example , morphological type .
the correlation between galaxy type and clustering amplitude was first seen by @xcite as a difference in the clustering slope of morphologically - selected elliptical and spiral galaxies , with ellipticals having a steeper slope .
this implied that ellipticals are more tightly clustered than spirals and that a neighboring galaxy of an elliptical was more likely be to another elliptical .
this was confirmed by @xcite who found a tight relation between morphology and local galaxy density in clusters of galaxies , with ellipticals preferentially located near the centers of clusters .
the color dependence of galaxy clustering has been measured more precisely in subsequent surveys of the local universe ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , with the most statistically - significant results provided by the latest large surveys , the 2df galaxy redshift survey ( 2dfgrs ) and the sloan digital sky survey ( sdss ) @xcite .
table 1 provides an overview of clustering results for color- and spectral type- selected samples from 2dfgrs and sdss as well as recent intermediate redshift surveys . in the 2dfgrs
, galaxies are classified into early- and late - type using a spectral analysis instead of using restframe colors @xcite .
@xcite find that early - type galaxies are much more strongly clustered than late - type , with a relative bias within spheres of 8 radius of @xmath6 .
@xcite investigate the luminosity dependence of clustering of early- and late - type galaxies in the 2df data and find that while there is a strong difference in the relative mix of early- and late - types as a function of luminosity , at @xmath7 this mix is not driving the overall increase of clustering for bright galaxies .
both galaxy types display a clustering increase at higher luminosities , with a constant relative bias between early- and late - types as a function of magnitude at @xmath7 .
these trends are confirmed by @xcite and @xcite using sdss data .
these authors find that both red and blue populations show a significant increase in the correlation length with luminosity , though within the blue cloud there is no significant change in the slope of for the brighter galaxies .
@xcite find that within the red population , however , the faintest galaxies ( @xmath8 ) have a steeper slope , which is likely due to a higher fraction of these galaxies being satellites of central galaxies in groups and clusters @xcite . color dependence of galaxy clustering has been detected in a variety of intermediate redshift ( @xmath9 ) surveys ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the most robust results to date at @xmath0 are from the vimos - vlt deep survey ( vvds ; * ? ? ?
* ) and the deep2 galaxy redshift survey @xcite .
using vvds data , drawing from a total sample of 6500 galaxies at @xmath10 over @xmath11 deg@xmath12 , @xcite measure the clustering of red and blue galaxies , dividing the two classes using the observed color bimodality .
they find that red galaxies have a higher correlation length and steeper slope than blue galaxies at @xmath13 , and that the relative bias of red to blue galaxies does not depend significantly on luminosity @xcite .
these results are consistent with those of @xcite , who use an early deep2 data sample of 2200 galaxies at @xmath14 in @xmath15 deg@xmath12 to determine that galaxies redward of the median color of the sample , @xmath16 , have a significantly higher correlation length and steeper slope than bluer galaxies . unfortunately , previous studies at intermediate redshift covered relatively small volumes and are dominated by large cosmic variance errors .
here we improve upon earlier @xmath0 deep2 results using the completed deep2 data set and volume - limited samples .
we measure galaxy clustering in several restframe color bins , with a particular emphasis on ` green valley ' galaxies at the intersection between the red / early- and blue / late - type populations , and as a function of luminosity within the red and blue galaxy populations separately .
the correlation between galaxy type and clustering strength is also reflected in measurements of the ` environment ' or local galaxy overdensity around objects of a given class .
locally , @xcite find a strong relation between local overdensity and galaxy color in sdss data , on scales of both 1 and 8 , with red galaxies favoring more overdense regions .
they find little dependence of overdensity on luminosity for galaxies in the blue cloud ( though the brightest blue galaxies do appear to be in more overdense regions ) and a strong dependence on luminosity for red galaxies locally , with both the brightest and faintest red galaxies being in denser environments than intermediate luminosity red galaxies . the lack of a stronger luminosity - overdensity relation for blue cloud galaxies may be somewhat at odds with the clustering results from 2df and sdss , which find a steeper , roughly linear trend of relative bias with magnitude , in that brighter blue galaxies are significantly more clustered than fainter blue galaxies @xcite .
@xcite , however , also study the clustering of blue galaxies in sdss and do not find a significant dependence on the relative bias with luminosity , in the same magnitude range as these other local studies , in contrast to the results of @xcite .
galaxy overdensity as a function of color and luminosity was recently measured for deep2 galaxies by @xcite , who find a similar dependence on color as is seen locally , and a dependence on luminosity for the brightest of the blue galaxies , which lie in overdense regions .
no dependence on luminosity was found for galaxies on the red sequence , in the luminosity range probed by deep2 .
we expand upon that study here by measuring the clustering as a function of scale ( instead of using a scale - averaged overdensity ) using a statistic that may more readily be compared with simulations , as it does not depend upon the bias of the tracer sample used , as is the case with overdensities .
the two - point correlation function results presented here will be used for hod modeling of the galaxy population at @xmath0 as a function of galaxy type in a future paper .
significant deviations in the two - point correlation function from a power - law form have recently been detected on scales @xmath17 for both local and high redshift populations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
. these departures from a power law are naturally explained in the hod framework , which provides a statistical understanding of the relation between galaxies and their dark matter halos , as the transition between pairs of galaxies within the same halo ( the ` one - halo ' term ) on small scales and galaxies in different halos ( the ` two - halo ' term ) on larger scales @xcite .
small - scale deviations from a power law are expected to be more significant at @xmath18 , where the exponential tail of the mass function moves to smaller masses , so that at a given mass scale the slope of the mass function becomes steeper @xcite .
observationally , deviations are also seen at a given redshift for the brightest galaxy samples , with @xmath7 .
this is likely due to the satellite fraction decreasing for brighter samples ; brighter galaxies are more likely to be central galaxies than satellite galaxies @xcite , which leads to a steeper one - halo term ( tinker et al , in prep . ) as more pairs are central - satellite pairs rather than satellite - satellite pairs and have , on average , a smaller separation
. departures of the two - point correlation function for deep2 galaxies from a power law as a function of luminosity are presented in @xcite ; here we perform a similar analysis for galaxies in deep2 as a function of color .
given that the galaxy correlation function is not a perfect power law , quoted and @xmath3 fits clearly do not fully capture all of the information present in clustering measures .
however , they are still useful as an approximate measure of the amplitude and shape of the correlation function on scales of @xmath19 and facilitate simple comparisons between observational results and models . as such , we provide and @xmath3 fits here but also discuss deviations from a power - law form for the data samples investigated .
the outline of the paper is as follows : 2 briefly describes the deep2 survey and defines the data samples used here . in
3 we discuss the methods used in this paper to measure the two - point correlation function , infer the real - space correlation length and slope , and correct for observational biases .
4 presents our results on the color dependence of galaxy clustering at @xmath0 . in 5
we compare our results with the semi - analytic models of @xcite applied to the millenium run dark matter simulation , and we conclude in 6 .
the deep2 galaxy redshift survey is a recently - completed project using the deimos spectrograph @xcite on the 10 m keck ii telescope to survey optically - selected galaxies at @xmath20 in a comoving volume of approximately 5@xmath2110@xmath22 @xmath23 mpc@xmath24 . using @xmath25 hour exposure times ,
the survey has measured high - confidence ( @xmath26% , quality flag @xmath27 or @xmath28 ) redshifts for @xmath29 galaxies in the redshift range @xmath30 to a limiting magnitude of @xmath31 .
the survey covers three deg@xmath12 of the sky over four widely separated fields to limit the impact of cosmic variance .
each field is composed of 2 - 4 ` pointings ' , each of which covers an area of @xmath11 by @xmath32 deg , corresponding to the area covered by the cfht12k camera ; there are 10 pointings in total .
the deep2 spectra have high resolution ( @xmath33 ) , and rms redshift errors ( determined from repeated observations ) are @xmath34 km s@xmath35 .
details of the deep2 observations , catalog construction and data reduction can be found in @xcite , @xcite , @xcite and @xcite .
k - corrections , absolute @xmath36 magnitudes and restframe @xmath37 colors have been derived as described in @xcite .
we do not include luminosity evolution in the k - corrections for @xmath36 .
absolute magnitudes given in this paper are in the ab system and are @xmath38 @xmath39 with @xmath40 , which for the remainder of the paper we simply denote as @xmath36 . to convert measured redshifts to comoving distances along the line of sight , we assume a flat cosmology with @xmath41 and @xmath42 . we define @xmath43 and quote correlation lengths , , in comoving . from the full flux - limited deep2 data set
we define a variety of volume - limited subsamples , each corresponding to some range in galaxy @xmath37 color and @xmath36 magnitude .
a volume - limited sample is not subject to redshift - dependent selection effects , i.e. , a galaxy of given rest - frame properties will be within such a sample regardless of its redshift . while volume - limited samples from a given survey will include fewer galaxies than flux - limited samples , they are much easier to interpret and compare with theoretical models , and are therefore used here .
details of each sample are given in table 2 , and cuts in color , magnitude and redshift are shown in figure [ mbz ] .
following @xcite , we define red and blue galaxies using the observed color bimodality in deep2 with the following tilted cut in color - magnitude space ( in ab magnitudes ) : @xmath44 we do not allow this color cut to evolve with redshift within the redshift range used here .
the color samples are shown in figure [ mbz ] , where the upper panels are color - magnitude diagrams for all deep2 galaxies with @xmath45 .
main _ red and blue samples are defined as those galaxies lying redward or blueward of the line defined in eqn .
1 , with @xmath46 ( dashed lines in the upper left panel of figure [ mbz ] ) and in the redshift range @xmath47 for red galaxies and @xmath48 for blue galaxies . while the redshift ranges are not identical for the main red and blue samples , differences in their clustering properties are due to the color differences and not the redshift ranges , as we have checked that using the same redshift range for both does not change the results . using similar redshift ranges also minimizes the effects of cosmic variance , as any density fluctuations will ( to first order ) be in common between the red and blue samples .
we define green galaxies as those within @xmath49 of the red / blue dividing line ( shown as dot - dash lines in figure [ mbz ] ) in the redshift range @xmath50 .
the green galaxy sample contains a subset of galaxies from each of the red and blue galaxy samples ; it is not defined as a distinct set of objects , which would require removing green galaxies from the main red and blue samples .
we further divide the red and blue samples into finer color bins : red galaxies are divided into redder and bluer halves at @xmath51 , while blue galaxies are divided into three color samples at @xmath52 and @xmath53 .
these color cuts are shown in the upper right panel of figure [ mbz ] .
these finer color cuts are independent of magnitude and do not have the a tilt in the color - magnitude diagram , unlike the red / blue dividing line
. our results do not change significantly if we instead define all cuts to be dependent on both color and magnitude .
the magnitude range we sample here is not wide , and the number of galaxies near these dividing lines is small enough that it does not significantly affect our results . within the red and blue galaxy populations we also construct subsamples as a function of threshold luminosity .
for the brighter luminosity samples we allow the upper redshift limit to increase with increasing luminosity , so as to include more galaxies and reduce measurement errors .
the lower panels of figure [ mbz ] show @xmath36 as a function of redshift for red ( left ) and blue ( right ) galaxies .
the dashed lines indicate the various luminosity samples defined in table 2 .
however , as illustrated in figure 2 of @xcite , the @xmath54 deep2 target selection limit defines a restframe color - dependent selection which changes as a function of redshift and which is not evident in figure [ mbz ] .
figure 2 of @xcite contains @xmath37-@xmath36 color - magnitude diagrams of deep2 galaxies in narrow redshift bins of @xmath55 .
the redshift - dependent selection is clearly evident , with the result being that red galaxies are not included in the survey at a given @xmath36 which is complete for blue galaxies .
this results in a lower upper redshift limit of the red samples used here compared to the blue samples of the same luminosity , though the mean redshift of the samples does not vary by more than @xmath56 .
number densities for each sample are given in table 2 ; these are derived from the deep2 luminosity function split by color @xcite . to account for redshift failures we use ` minimal ' weighting for the blue galaxies and ` average ' weighting for the red galaxies ( see @xcite for details ) .
we estimate the number density for the green galaxy sample by multiplying the number densities for the main red and blue samples by the percentage of galaxies in each color that are in the green valley in the redshift range @xmath57 ( 12.2% of blue galaxies and 21.6% of red galaxies ) . the dominant error in the number density estimate derived from the luminosity function is the error on @xmath58 ; using the errors quoted in @xcite , the fractional errors on the number densities for the main red and blue samples ( with @xmath59 ) are @xmath60 for blue galaxies and @xmath61 for red galaxies . both locally and to
at least @xmath62 galaxies are found to be bimodal not only in terms of their restframe color distribution ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) but also in morphological and spectral type ( e.g. , * ? ? ?
* ; * ? ? ?
locally , there are strong correlations between morphology , spectral type and restframe color , with some scatter that likely reflects both intrinsic variation and the difficulty of unequivocally determining morphological type . defining galaxy samples by spectral type in deep2 is not as straightforward as using restframe colors @xcite , due to the limited spectral range of the deep2 spectra , which is a consequence of working at high resolution .
we therefore define galaxy samples here by restframe color and luminosity ; however , we show in figure [ coadd ] coadded rest - frame spectra for galaxies in each of our color samples in order to investigate and illustrate the physical nature of the optical emission for galaxies in each sample .
these coadded spectra demonstrate that our color divisions correspond directly to classical spectral types .
the coadditions correct for the relative probability that each galaxy was selecteding for observation by deep2 , but otherwise all objects are given equal weight . to maximize signal - to - noise and robustness , pixels affected by night sky emission
are deweighted and those affected by instrumental artifacts are removed .
the top row of figure [ coadd ] includes the coadded spectra for the main blue and red galaxy samples .
the blue galaxies display strong emission lines , primarily from star formation ( [ oii ] @xmath63 , h@xmath3 , h@xmath64 , [ oiii ] @xmath65 ) , along with weak ca h+k @xmath66 absorption and stronger balmer absorption features .
the main red galaxy sample has a small amount of [ oii ] and [ oiii ] emission , primarily from agn @xcite , as well as ca h+k absorption and weak balmer absorption features .
we also show coadded spectra for galaxy samples selected in finer color bins . within the blue cloud ,
the bluer galaxies have much stronger star - forming emission lines in equivalent width , while the redder galaxies have lower ew lines ; relative to the normalized continuum ; in luminosity the redder of the blue galaxies have stronger [ oii ] .
the green galaxies , composed of both blue and red galaxies , have lower ew star - forming lines than the blue galaxies .
the green galaxy spectrum does not appear to be a simple mix of the red and blue galaxy spectra , as the line ratios are different .
the green galaxy spectrum has a much higher [ oiii]/h@xmath64 ratio than the blue galaxy spectrum or the average of the ` blue : redder ' and ` red : bluer ' spectra ; this indicates a higher fraction of agn in the green population relative than in the blue cloud as a whole @xcite , and also shows that the green galaxies are not a simple average of the redder of the blue cloud galaxies and the bluer of the red sequence galaxies .
the high [ oiii]/h@xmath64 ratio in the green valley population is likely dominated by the red galaxies in the sample @xcite . within the red sequence ,
the bluer galaxies have a higher [ oiii]/h@xmath64 ratio , indicating more agn activity , and the redder galaxies have stronger ca h+k absorption features .
the two - point correlation function is defined as a measure of the excess probability @xmath67 ( above that for an unclustered distribution ) of finding a galaxy in a volume element @xmath68 at a separation @xmath69 from another randomly - chosen galaxy , @xmath70 dv,\ ] ] where @xmath71 is the mean number density of the galaxy sample in question @xcite . for each galaxy sample we construct a randomly - distributed catalog with the same overall sky coverage and redshift distribution as the data .
we then measure the two - point correlation function using the @xcite estimator , @xmath72,\ ] ] where @xmath73 , and @xmath74 are counts of pairs of galaxies ( as a function of separation ) in the data data , data random , and random random catalogs , and @xmath75 and @xmath76 are the mean number densities of galaxies in the data and random catalogs .
this estimator has been shown to perform as well as the hamilton estimator @xcite but is preferred here as it is relatively insensitive to the size of the random catalog and handles edge corrections well @xcite . to estimate the cross - correlation function between two galaxy samples , we measure the observed number of galaxies from a given sample around each galaxy in another sample as a function of distance , divided by the expected number of galaxies for a random distribution .
for this we use the simple estimator @xmath77 where @xmath78 are pairs of galaxies between the two samples and @xmath79 are galaxy - random pairs between one galaxy sample and the random catalog of the other sample , where the pair counts have been normalized by @xmath75 and @xmath76 .
the deep2 redshift success rate is @xmath8070% overall ( defined as the percentage of galaxies targeted for spectroscopy that have a well - determined redshift ) and is not entirely uniform across the survey ; some slitmasks are observed under better conditions than others and therefore yield a slightly higher completeness .
we only use regions of the survey with a redshift success rate @xmath8065% , and the spatially - varying success rate is taken into account in the window function , which is applied to the random catalog to ensure that it has the same spatial distribution as the survey . the two - dimensional window function of the deep2 data in the plane of the sky reflects the probability of observing a galaxy and takes into account the overall outline of the survey and the geometry of the overlapping slitmasks as well as vignetting in the deimos camera and gaps between the deimos ccds .
the varying redshift success is also taken into account such that regions of the sky with a higher completeness have a correspondingly higher number of random points .
this ensures that there is no bias introduced when computing correlation statistics .
we also mask the regions of both the random and mock catalogs where the photometric data have saturated stars or ccd defects .
redshift - space distortions due to peculiar velocities along the line of sight significantly affect estimates of . at small separations ,
random motions within a virialized overdensity cause an elongation along the line of sight ( `` fingers of god '' ) , while on large scales , coherent infall of galaxies into potential wells causes an apparent contraction of structure along the line - of - sight @xcite .
while these distortions can be used to uncover information about the underlying matter density and thermal motions of the galaxies , they complicate a measurement of the two - point correlation function in real space . in order to separate the effects of these redshift - space distortions and uncover the underlying real - space clustering properties , we measure @xmath81 in two dimensions , both perpendicular to and along the line of sight . following @xcite ,
we define vectors @xmath82 and @xmath83 to be the redshift - space positions of a pair of galaxies , @xmath84 to be the redshift - space separation ( @xmath85 ) , and @xmath86(@xmath87 to be the mean coordinate of the pair .
we then define the separation between the two galaxies along ( @xmath88 ) and across ( @xmath89 ) the line of sight as @xmath90 @xmath91 to estimate , we apply the @xcite estimator to pair counts subdivided over a grid in @xmath89 and @xmath88 .
while contains useful information about peculiar velocities , we also wish to measure the real - space correlation function , . to recover we use a projection of along the @xmath89 axis .
as redshift - space distortions affect only the line of sight component of , integrating over the @xmath88 direction leads to a statistic , which is independent of redshift - space distortions .
following @xcite , @xmath92 , \label{eqn}\ ] ] where @xmath93 is the real - space separation along the line of sight .
if is modelled as a power law , @xmath94 , then and @xmath3 can be readily extracted from the projected correlation function , , using an analytic solution to equation [ eqn ] : @xmath95 where @xmath96 is the usual gamma function . a power - law fit to will then recover and @xmath3 for the real - space correlation function , . in practice ,
is not expected to be a power law at very large scales ( @xmath97 ) , nor can we measure accurately to infinite separations as assumed in equation [ eqn ] . here
we integrate to @xmath98 , as becomes noisy at larger separations .
it is not appropriate to then apply equation [ powerlawwprp ] directly ; for power - law correlation functions , that equation becomes increasingly inaccurate at large @xmath89 ( failing significantly where @xmath99 ) .
instead , we must compare the observed to the integral of over separations within @xmath100 , as predicted for a given and @xmath3 .
this prediction is complicated by the presence of redshift - space distortions due to coherent infall of galaxies ( @xcite ; see @xcite and section 4.1 of @xcite for the relevant equations for correlation function analyses ) .
if @xmath100 were infinite , these distortions would have no effect , as they merely change the line - of - sight separations of pairs of galaxies ; but that is not the case here .
the strength of these distortions depends on @xmath101 , where @xmath102 is defined at the mean redshift of the sample ( not @xmath103 ) and @xmath104 is the linear bias between the clustering of galaxies and dark matter .
we therefore recover and @xmath3 as follows .
we first measure in the data by integrating to @xmath98 .
we then estimate for dark matter particles at the mean redshift of the data sample , using the publicly - available code of @xcite ( not including redshift - space distortions ) , where we integrate the dark matter to @xmath98 .
we then estimate the linear galaxy bias from the ratio of these quantities : @xmath105_{\rm gal}/[w_p]_{\rm dark matter}$ ] .
we use the average value of this @xmath104 over scales @xmath106 and use @xmath102 at the mean redshift of the sample ( for an lcdm model with @xmath107 ) , to provide an estimate of the redshift - space distortion parameter @xmath64 . using @xmath108 rather than 0.3
increases our recovered values by 1% and does not change @xmath3 .
we can then predict the observed for any point in this grid by integrating to @xmath100=20 .
the best - fit and @xmath3 values are those that minimize the @xmath109 difference between the data and the model prediction .
although our estimated bias , and hence the @xmath64 used in this procedure , may not perfectly match the bias corresponding to the best - fit values of and @xmath3 , the measured values prove to be insensitive to that estimate .
increasing the bias by 20% increases by 1% and decreases @xmath3 by 1% .
this method for recovering and @xmath3 assumes that is a power law only to a scale of @xmath100 and results in and @xmath3 values within a few percent of those obtained using equation [ powerlawwprp ] , for our value of @xmath98 .
deviations from equation [ powerlawwprp ] are significant only on larger scales , where @xmath110 .
we also consider an alternative method for recovering correlation functions in the next section .
errors on are calculated using the standard error across the 10 separate data pointings .
we do not use a full covariance matrix in the fits .
errors on and @xmath3 are derived from jacknife resampling of the separate pointings and therefore take into account the covariance among the @xmath89 bins . given that it becomes necessary to model the effects of the truncation of @xmath111 and include redshift - space distortions when comparing theory with observations , padmanabhan , white and eisenstein ( 2007 , hereafter pwe07 ) propose an alternative statistic , @xmath112 , which has better convergence properties to the real - space quantities .
this statistic is defined as @xmath113 where again @xmath89 and @xmath88 are transverse and line of sight directions , respectively .
the filter function , @xmath114 is chosen to be non - zero only for @xmath115 and to have zero integral ( when integrated against an @xmath116 measure ) .
we follow pwe07 and adopt @xmath117 with @xmath118 this statistic has the further advantages of being unbinned and having a well - localized kernel in @xmath119 , probing the correlation function at a scale @xmath120 .
we note that @xmath112 can be thought of as performing the abel inversion from @xmath121 to @xmath81 , while avoiding differentiating noisy data . in order to test the robustness of the methods presented above , below in section 4.1 we compute @xmath122 for the main red and blue galaxy samples to compare with the results obtained using the methods in section 3.2 .
to compute @xmath122 we use the @xcite estimator for the two - dimensional correlation function , and the techniques in pwe07 to rewrite the integrals in eqs .
[ eq : omegadef ] as weighted pair sums ; the line of sight integral is truncated at @xmath123 .
uncertainties in @xmath122 are estimated by jackknife resampling the ten deep2 pointings . since @xmath122 is unbinned
, one can choose arbitrarily finely spaced bins , with the disadvantage that adjacent bins become highly correlated .
we have verified explicitly that our fits for and @xmath3 given below are insensitive to the particular choice of binning we adopt .
it is straightforward to relate @xmath122 to the three - dimensional correlation function , @xmath81 . for a power - law correlation function , @xmath124 , and our choice of @xmath125 ,
this relation becomes @xmath126 when observing with multi - object slitmasks , the spectra of targets can not be allowed to overlap on the ccd array ; therefore , objects that lie near each other in the direction on the sky that is perpendicular to the wavelength direction on the ccd can not be simultaneously observed .
this results in undersampling the regions of the sky with the highest density of targets . to reduce the impact of this bias , adjacent
deep2 slitmasks are positioned approximately a half - mask width apart , giving each galaxy two chances to appear on a mask ; we also adaptively tile the slitmask locations to hold constant the number of targets per mask . in spite of these steps ,
the probability that a target is selected for spectroscopy is diminished by @xmath127% if the distance to its second nearest neighbor is less than 10 arcseconds ( for technical details of the slitmask design , see @xcite ) .
this introduces a roughly predictable systematic bias which leads to underestimating the correlation strength on small scales .
the objects that are in conflict are often not at the same redshift , however , such that the effect is not particularly large . to correct for this effect we use the mock catalogs of @xcite with galaxy colors added as described in @xcite .
we measure the ratio of both and in the mock catalogs between samples of galaxies from catalogs with and without the slitmask target selection algorithm applied .
we then multiply the measured and in the data by this ratio , which is a smooth function of scale and varies from @xmath128% on the smallest scales to 2% on scales greater than @xmath129 .
we tested mock catalog samples with identical redshift , luminosity and color ranges as each data sample used here and found no significant differences in the multiplicative corrections as a function of color or luminosity , such that the same correction derived from the largest sample is our best estimate for all other samples .
the effect of this slitmask correction is small ; it increases by 1.5% and @xmath3 by 2.5% .
we include an additional error due to this correction , which is added in quadrature to the error on .
this error is estimated from the variance on the correction in the mock catalogs and is a function of scale , varying from 10% at @xmath130 to 2% on scales @xmath131 .
we first compare the main blue and red galaxy samples , both of which have @xmath46 .
the top left panels in figure [ xisp_all ] show for these two samples , with contours indicating constant probability , where the dark line is @xmath132 .
there are several clear trends present in these diagrams ; red galaxies are seen to be more clustered than blue galaxies ( a given @xmath81 is seen at larger separations ) and show much stronger fingers of god , as seen in the elongation along the @xmath88 direction at small @xmath89 separations ( @xmath1331 ) .
both of these results reflect the fact that a color - density relation is in place at @xmath0 , with red galaxies residing more in overdense regions such as groups and clusters than blue galaxies .
infall on large scales is plainly seen in these diagrams for blue galaxies , as indicated by the flattening of the contours along the line of sight .
coherent infall may be present for red galaxies as well , though it is harder to detect due to the larger fingers of god ( see section 4.9 for more details ) .
redshift - space distortions are significant out to @xmath134 , such that it is important to not use the usual power - law estimator ( equation [ powerlawwprp ] ) to infer and @xmath3 from , as this equation assumes that one has integrated to @xmath135 .
the top left panel of figure [ wprp_all ] shows the projected correlation function , , for the main red and blue samples .
red galaxies exhibit a significantly steeper slope than blue galaxies , and both samples are well approximated by a power law within the error bars .
power - law fits are given in table 2 , fitting the data on scales @xmath136 ; the smallest @xmath89 bin ( 0.05 ) is not used in these fits due to the rise on small scales which is seen in some samples ( see section 4.3 for further discussion of deviations on small scales ) . including the smallest bin in these fits , or fitting only on scales larger than @xmath129 , does not change the results for the main blue and red galaxy samples . the correlation length for red galaxies is found to be significantly higher than for blue galaxies : @xmath137 compared to @xmath138 .
these values for at @xmath0 are somewhat lower than what is found for local samples ( table 1 ) ; however , has only increased by @xmath139515% by @xmath140 for blue galaxies and @xmath1391020% for red galaxies .
furthermore , there has been very little evolution in the slope of the correlation function for either red or blue galaxies since @xmath0 .
we discuss the implications of this in section 6 .
our values for blue and red galaxies are somewhat higher than those reported by @xcite and @xcite , though generally within the large errors in those findings .
@xcite use early data from a portion of one deep2 field and @xcite use the vvds deep survey first epoch data ; both of these samples are roughly 1/6 of the dataset used here , and neither @xcite nor @xcite constructs volume - limited subsamples . for blue galaxies ,
our value found here is 2@xmath141 and 5@xmath141 higher than @xcite and @xcite , while for red galaxies our value is 1@xmath141 and 2@xmath141 higher , respectively , where the errors have been added in quadrature .
@xcite measure the clustering of red galaxies in the noao deep wide - field survey , using photometric redshifts .
they find at @xmath142 that @xmath143 for red galaxies with @xmath144 ( roughly @xmath145 ) , which is 2@xmath141 higher than our result .
@xcite measure the clustering of uv - selected galaxies , again using photometric redshifts , and find that at @xmath146 @xmath147 and @xmath148 . their clustering scale - length is 2@xmath141 higher than what we find here for star - forming galaxies .
we also use the estimator of pwe07 , as described in section 3.3 , to measure @xmath122 for the main blue and red samples ; the results are shown in fig . [ omega ] .
we have not attempted to correct for slitmask effects in the @xmath122 estimator ; these effects are small , as discussed above . using eqn .
[ eq : poweromega ] , we fit for @xmath149 and @xmath3 for both galaxy samples .
the resulting power - law fits are plotted in fig .
[ omega ] ( solid lines ) , and are compared with the fits obtained from fitting to @xmath121 ( dashed lines ) .
the two fits are clearly consistent with each other , given the measurement errors .
for the blue galaxy sample , we obtain @xmath150 and @xmath151 , while for the red galaxies , we find @xmath152 and @xmath153 .
these and @xmath3 values are consistent with the estimates derived above from fitting to @xmath121 . in fact , because the correlation coefficient between and @xmath3 is large and negative ( @xmath154 ) , difference between the results of the two estimators are much less significant than the projected errors imply .
the error bars on @xmath149 and @xmath3 are larger using the @xmath122 estimator , which is not surprising given that @xmath122 is an inversion of @xmath121 to a localized integral of @xmath119 .
in addition to measuring the auto - correlation of blue and red galaxies separately , the cross - correlation between the two samples can be used to further understand the spatial relationship between blue and red galaxies .
the cross - correlation between two samples measures the clustering of one type of object around the other and provides information on the mixing of the populations . on large , linear scales , in the two - halo regime
, blue and red galaxies should be well - mixed and trace the same overall large - scale structure , such that the cross - correlation should equal the geometric mean of the auto - correlations of each sample . on small scales , however , within the one - halo term , deviations from the geometric mean encode information about the differences in the halos the two samples populate .
for example , if some halos contain almost exclusively red galaxies and others only blue galaxies , then the cross - correlation function would fall below the geometric mean .
the blue - red galaxy cross - correlation function can therefore be used to distinguish between different star formation quenching scenarios , as it traces the distribution of quiescent and actively star - forming galaxies around each other .
here we measure the cross - correlation of the main blue and red galaxy samples to investigate in particular the mixing of blue and red galaxies on small , one - halo - dominated , scales ( @xmath155 ) .
the two - dimensional cross - correlation function is shown in the lower left panel of figure [ xisp_all ] and the projected correlation function , , is shown in the upper right panel of figure [ wprp_all ] .
the cross - correlation is intermediate between the auto - correlation of blue and red galaxies .
the best fit scale length is @xmath156 and slope is @xmath157 .
fingers of god are clearly seen in , though the coherent infall on large scales appears to be more similar to what is seen for the red sample than the blue sample .
figure [ cross_powerlaw ] shows for the main blue and red galaxy samples and the red - blue cross - correlation function divided by a reference power law with @xmath158 and @xmath159 .
also plotted are deviations from this power law for the geometric mean : @xmath160 while the cross - correlation ( dot - dash line ) is consistent with the geometric mean ( dotted line ) within the error bars , on scales @xmath155 ( the one - halo term ) the cross - correlation is closer to the blue than the red .
this reflects a deficit of blue galaxies near red galaxies , which could be due to the centers of groups being occupied preferentially by red galaxies .
this is consistent with the observed cross - correlation between deep2 groups and galaxies @xcite , which is higher for red galaxies than blue galaxies , again indicating that red galaxies tend to be at the centers of groups . a similar effect has been seen at @xmath161 in the sdss red - blue cross - correlation function .
both @xcite and @xcite find that the red - blue cross - correlation amplitude is lower than the geometric mean of the red and blue clustering on small scales .
@xcite measure the cross - correlation as a function of luminosity and find a suppression relative to the geometric mean for faint red and faint blue galaxies , but not for brighter samples .
@xcite find a similar effect in their sdss group catalog , which they term ` galactic conformity ' , in which the colors of satellite galaxies are correlated with the colors of their central galaxies ; i.e. , red satellite galaxies tend to be around red central galaxies , and the same is seen for blue galaxies .
the luminosity dependence of clustering in the overall deep2 galaxy sample was presented in @xcite ; here we expand upon that study by computing the luminosity dependence of clustering within the blue and red galaxy populations separately .
the median @xmath36 of each sample is given in table 2 .
for comparison , at @xmath162 , @xmath163 is @xmath164 for blue galaxies and @xmath165 for red galaxies ( from willmer et al .
2006 , where we use the ` minimal ' weighting for blue galaxies and ` average ' weighting for red galaxies ) .
the main blue sample has a median @xmath36 value near @xmath163 , while the blue luminosity subsamples range from @xmath166 magnitudes fainter than @xmath163 to @xmath167 magnitudes brighter than @xmath163 .
the main red sample is @xmath168 magnitudes brighter than @xmath163 ( in the median ) for red galaxies , and the red luminosity subsamples range from @xmath169 to @xmath170 magnitude brighter than @xmath163
. figures [ xisp_blue_lum ] and [ xisp_red_lum ] show for four luminosity threshold samples for blue and red galaxies . within the blue population ,
there is almost no difference between the faintest two samples , while the brightest samples may show somewhat stronger fingers of god .
all of the blue luminosity samples show coherent infall on large scales .
the red shows almost no difference between any of the bins ; the brightest is noisy , however , due to the small sample size .
any correlation between the strength of the small - scale redshift - space distortion and luminosity within the red galaxy population , for the luminosity range that we probe here , is too weak for us to detect .
@xcite find that the pairwise velocity dispersion ( see section 4.8 ) , which quantifies the amplitude of the fingers of god , increases with luminosity for blue galaxies in sdss , while red galaxies do not show any luminosity dependence , qualitatively similar to our results at @xmath0 . the projected correlation function , , is shown in figure [ wprp_lum ] for blue subsamples ( top row ) and red subsamples ( bottom row ) as a function of luminosity .
again there is little difference in clustering between the fainter blue samples ( upper left panel ) ; however displays a significant rise on small scales for the brighter blue samples ( upper right panel ) .
the red samples show no corresponding rise on small scales for the brighter samples ( lower right panel ) .
deviations from a power law for the brighter samples are shown in figure [ wprp_lum_powerlaw ] ; on scales @xmath171 there is a clear rise in the correlation function of blue galaxies .
a similar increase on small scales was seen for all deep2 galaxies in @xcite .
this is likely due to brighter galaxies having , on average , a higher probability of being a central galaxy rather than a satellite galaxy @xcite , which leads to a relative rise in the one - halo term compared to the two - halo term ( tinker et al , in prep . ) .
there may also be a contribution from interaction - driven starbursts ( e.g. , * ? ? ?
the lack of a strong rise on small scales in the brighter red samples likely results from red deep2 galaxies having a higher satellite fraction than blue galaxies , at a given luminosity .
we note that while the brighter blue galaxies are more likely to be central galaxies , they are not necessarily in more massive halos , as shown by the lack of significant dependence of on luminosity within the blue cloud .
power - law fits are given in table 2 where , as above , the scales used to fit a power law are @xmath136 . fitting on scales @xmath172 instead
does not significantly change the results .
the results are shown in the upper panels of figure [ r0_gam_lum ] .
there is no significant dependence in the correlation length , , or slope , @xmath3 , on luminosity for either blue or red galaxies , within the error bars .
@xcite use local galaxy overdensity measures to determine the environment of deep2 galaxies and conclude that within the luminosity range sampled by deep2 there is no significant dependence on luminosity for red galaxies , although there is for blue galaxies , in that brighter blue galaxies are in more overdense regions . within the luminosity range probed here ,
the relative bias of blue galaxies is found by @xcite to increase by @xmath13930% , which is steeper than the trend found here using the correlation function .
the brighter blue samples do show an increase in the small - scale clustering amplitude relative to the fainter blue samples , however , which is not reflected in the measured @xmath173 values .
@xcite also measure local overdensity at @xmath174 in vvds data and find that the fraction of red galaxies at a given overdensity depends significantly on luminosity .
we do not find such a trend here with our larger dataset .
the lower panels of figure [ r0_gam_lum ] compare the clustering scale - length of red and blue deep2 galaxies as a function of @xmath175 with local results from 2df @xcite for passive and active galaxies , defined using spectral types , and from sdss @xcite , using restframe colors .
for this comparison we use @xmath176 for all galaxies at @xmath162 @xcite and @xmath177 and @xmath178 for all galaxies at @xmath161 @xcite . at @xmath161 , @xcite all
detect significant luminosity dependence in the clustering of both blue and red galaxies , while @xcite detect luminosity dependence for red galaxies only .
@xcite detect significant luminosity dependence in the clustering of both passive and active galaxies as well .
this luminosity dependence for local blue and red galaxies is significant at @xmath7 but not at @xmath179 ; there is also a significant increase in the clustering for faint red galaxies , but that is below the luminosity range that we probe here .
the deep2 sample does not probe @xmath180 , such that we can not test the luminosity dependence at large @xmath175 . within the range that we do probe
, it does not appear that either blue or red galaxies have a detectable luminosity dependence in their autocorrelation properties .
this is consistent with local results in the same @xmath175 range .
our result that the clustering amplitude does not strongly depend on luminosity for red galaxies at @xmath0 is at odds with @xcite , who find a significant luminosity dependence in the clustering of red galaxies at @xmath181 , for samples between @xmath182 and @xmath183 ( in vega magnitudes ) , using photometric redshift data .
@xmath184 for the @xcite sample is @xmath185 ( m. brown , private communication ) .
their value of @xmath173 at @xmath186 for @xmath187 red galaxies is @xmath188 , which is only somewhat higher than that found here ( @xmath137 at @xmath189 ) or at @xmath140 by @xcite ( @xmath190 ) .
however , they find a steep trend in @xmath173 with @xmath191 , with @xmath173 ranging from @xmath192 to @xmath193 for @xmath194 , which is not found here or locally .
no value of @xmath184 would lead to a consistent trend of @xmath173 with luminosity when compared with our results or local results from @xcite or @xcite .
estimating from an angular correlation function , as done in @xcite , requires very accurate knowledge of the redshift distribution ; this may account for the discrepancy with the results from spectroscopic surveys .
@xcite find that within the full deep2 galaxy sample the clustering strength depends strongly on luminosity .
this can still occur even if there is no luminosity dependence for red or blue samples separately , as the red galaxy fraction is a function of luminosity ; the higher prevalence of red galaxies in the brighter samples leads to increasing as @xmath36 decreases .
the clear bimodality both locally and at @xmath62 in galaxy properties , whether defined by restframe color , morphology or spectral type , raises the question of how these two main types arise and what , if any , evolutionary connection exists between them .
the observed buildup of galaxies on the red sequence since @xmath62 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) suggests that blue galaxies are moving to the red sequence with time , though the details of this transition are not clear .
the general outline , however , is that star formation begins to shut down or be quenched in blue galaxies , after which they passively evolve onto the red sequence .
the star formation quenching process(es ) involved may be causally related to the environment of the galaxy ( e.g. , ram pressure or tidal stripping of gas @xcite ) , such that galaxies in higher density regions end their star forming phase more readily , thus creating a color - density relation .
alternatively , the process may be inherent to the galaxy and depend on its age or stellar or halo mass , in which case the process may be correlated with environment but not caused by it @xcite .
to elucidate what is causing this change , we can investigate the clustering properties of galaxies located at the transition region between these populations . here
we study ` green ' galaxies located near the minimum of the observed color bimodality as defined in 2.2 .
the green galaxy , shown in the upper right panel of figure [ xisp_all ] , exhibits an intermediate clustering amplitude between the blue and red galaxy samples , as well as intermediate - strength fingers of god .
the green population displays somewhat similar infall on large scales as blue galaxies , though the contours are noisy on large scales due to the relatively small sample size .
the projected correlation function for green galaxies is shown in the lower left panel of figure [ wprp_all ] , plotted with for red and blue galaxies for comparison . on large scales ( @xmath131 )
the green galaxies have a clustering amplitude similar to that of red galaxies , while on small scales the amplitude is closer to that of blue galaxies .
the large - scale agreement with the red galaxy population implies that green galaxies reside in or on the outskirts of halos of similar mass as red galaxies , as they have the same clustering for the two - halo term .
green galaxies are therefore generally in the same overdense regions as red galaxies . however , the one - halo term for green galaxies is lower than for red galaxies , which likely reflects a lower radial concentration of green galaxies within their parent halos .
the overall picture is consistent with galaxies having green colors while on the outskirts of halos and becoming redder as they reach the center of the overdensity .
these results point towards green galaxies being a distinct population and not a simple mix of red and blue galaxies . to investigate these trends further ,
we compute the cross - correlation between the green galaxy sample and the main blue and red galaxy samples .
the diagrams ( lower right panels of figure [ xisp_all ] ) show more significant infall on large scales in the green - blue cross - correlation than the green - red cross - correlation ( comparing , for example , the @xmath132 contours ) .
the green - red cross - correlation is very similar to the blue - red cross - correlation ( though with smaller fingers of god ) , indicating that green galaxies have comparable kinematics to blue galaxies .
there are also smaller fingers of god in the green - red cross - correlation than in the red auto - correlation .
these results are consistent with green galaxies residing in the same overdense regions as red galaxies , on average , but with a lower probability of being at the cores of those overdensities , compared to red galaxies .
this would result in green galaxies having smaller fingers of god than red galaxies , and a lower auto - correlation on small scales .
it is possible that green galaxies reside on the outskirts of the same halos that red galaxies occupy but are still falling in towards the centers of the halos .
the bottom right panel of figure [ wprp_all ] shows the projected cross - correlation functions between green and blue or red galaxies .
the green - red galaxy clustering amplitude is higher than the green - blue galaxy amplitude on all scales . on scales
@xmath106 the green - red sample is @xmath195% more clustered . on small scales , dominated by the one - halo term ,
the green - red amplitude is a few times higher than the green - blue amplitude : a factor of @xmath196 times higher on scales @xmath11 and a factor of @xmath197 times higher on scales @xmath198 .
one would therefore be _ much _ more likely to find a red close neighbor ( or satellite ) galaxy to a green galaxy than a blue neighbor , and vice versa , if red and blue galaxy samples were of equal size .
another way to frame these results is that a green galaxy is less likely to be a central galaxy than a red galaxy of similar luminosity , but more likely than a comparable blue galaxy .
we note that the green galaxy likely can not be fit with an hod model that assumes that the radial distribution of galaxies follows that of dark matter particles or subhalos .
the green galaxy radial distribution is clearly different from that of either red or blue galaxies , and most hod models do not allow the radial distribution to vary for different galaxy types .
we next divide the blue and red galaxy samples into finer color bins to investigate the strength of clustering as a function of color _ within _ each population .
figure [ r0_gam_color ] shows and @xmath3 as a function of color for both the main blue and red samples ( triangles ) , the green sample , and for finer color bins .
as explained in section 2.2 , the finer color bins are not all independent ; this is shown in the figure with a dotted horizontal line showing the color range for each point ( for clarity the color ranges for the main blue and red samples are omitted ) .
in particular , the green galaxy sample contains subsamples of both blue and red galaxies . within the red population ( where we have two independent samples ) neither or @xmath3 depend significantly on color
; across the red sequence we detect no change in the clustering properties of galaxies . within the blue population ,
however , there is a strong dependence ; redder galaxies are more clustered than the bluer galaxies .
there is no corresponding trend seen in the slope of the correlation function .
the abrupt change in slope between blue and red galaxies likely reflects the different radial concentrations and satellite fractions of the two populations and may indicate that star formation quenching is an effect associated with the halo properties of galaxies .
interestingly , the green galaxy is found to be the same as the red galaxy ( within 1 @xmath141 ) , while the green galaxy @xmath3 is the same as for blue galaxies , not red galaxies .
this is consistent with the interpretation in the previous section that green galaxies reside in similar mass halos as red galaxies , and are found generally in overdense regions , but are less concentrated within their host halos than red galaxies .
they apparently do not populate the centers of overdensities .
as discussed earlier , luminosity function studies find that a significant fraction of the progenitors of local galaxies on the red sequence must have been in the blue cloud at @xmath0 @xcite .
our clustering results imply that the galaxies that migrate to the red sequence from @xmath62 to @xmath103 are amongst the redder of the blue galaxies , and the green galaxies at @xmath62 , which are already as clustered as galaxies on the red sequence . these trends of finding a color - density relation within the blue cloud but not within the red sequence are similar to what has been locally using galaxy environment measures ( e.g. , * ?
* ; * ? ? ?
there are no other published @xmath0 clustering results with such fine color bins to compare to , and we defer a comparison with the deep2 environment results of @xcite to section 4.7 .
to facilitate comparisons to published clustering results measured at other redshifts or with different statistics than the two - point correlation function , we calculate the relative bias between various galaxy color samples .
we define the relative bias as the square root of the ratio of for two samples . as
the relative bias is scale - dependent , we calculate the mean relative bias over two scales , @xmath199 ( all scales ) and @xmath200 ( large scales ) , which have mean scales of @xmath201 and @xmath202 , respectively .
the relative bias of red to blue galaxies is @xmath203 on large scales and @xmath204 on all scales , reflecting the rise in the relative bias on small scales , @xmath155 .
we do not find any dependence on luminosity for the relative bias between red and blue galaxies .
this is in contrast to the results of @xcite , who measure the color - density relation at @xmath174 in vvds data and find that the color - density relation steepens for brighter galaxies .
no such trend is detected here .
the lack of luminosity dependence in the relative bias found here could indicate that the ratio of the satellite fraction in red versus blue galaxy samples does not depend on luminosity .
@xcite show that for central galaxies there is no color - density relation ; this relation only exists for satellite galaxies .
this implies that the relative bias of red to blue galaxies should be a function of luminosity if the central to satellite galaxy fraction changes with luminosity differentially for red and blue galaxies .
the fact that the data show no dependence in the relative bias on luminosity could imply that the ratio of the satellite fraction in red versus blue galaxy samples is fixed and does not depend on luminosity .
the relative bias of red to green galaxies is @xmath205 on large scales and @xmath206 on all scales .
given that red and green galaxies have different correlation slopes and redshift - space distortions , the relative bias does not fully reflect the differences between these populations .
it does , however , indicate that when two - halo - dominated scales are included , on average the clustering amplitude of green galaxies is perfectly consistent with that of red galaxies and is not lower .
the relative bias of green to blue galaxies is found to be @xmath207 on both large and all scales , with no scale - dependence .
the relative bias found here between red and blue galaxies at @xmath0 is very similar to what is measured at lower redshift ( see table 1 ; @xmath208 is @xmath209 % lower at @xmath0 ) .
@xcite compare the clustering of early and late type galaxies ( defined by spectral type ) and find a relative bias of @xmath210 integrated to scales of @xmath211 , while @xcite find a relative bias of @xmath212 between red and blue galaxies .
@xcite find using sdss data that the relative bias of red to blue galaxies is @xmath213 , but is highly scale - dependent .
this agreement with lower redshifts implies that the dependence of color on density at @xmath62 is as steep as it is locally ; the physical mechanisms responsible for the color - density relation are just as effective before @xmath62 as they are after .
the color - density relation was therefore established before @xmath62 and subsequent changes may be absolute ( in terms of the zeropoint of the relation ) but not relative ( which would change the slope ) , within the errors . as discussed in @xcite , given the low abundance of rich clusters at @xmath62 , the color - density relation seen in the deep2 data can not be the result of cluster - specific physics such as ram pressure stripping of gas or galaxy harassment ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . for a given cosmology we can also compute the _
absolute _ galaxy bias relative to the underlying dark matter density field using results from n - body simulations .
we use the power spectrum of @xcite to estimate the dark matter clustering amplitude at the same redshift as the data , for a cosmology with @xmath214 , and @xmath215 .
this code is an analytic fit motivated by the halo model ; we include a 5% error on the dark matter at each scale in our calculation of the absolute bias to reflect the uncertainty in the fit .
the resulting absolute bias values averaged on scales @xmath106 are listed for each color and magnitude sample in table 2 . for @xmath216
the large - scale bias is 13% higher , while for @xmath217 it is 10% lower .
note that the absolute bias and relative bias are measured on different scales .
we find that the absolute bias is @xmath218 for red galaxies and @xmath219 for blue galaxies , both with @xmath59 .
this is consistent with @xcite , who find that red galaxies with @xmath59 at @xmath220 have an absolute bias of @xmath221 and blue galaxies have @xmath222 , measured on scales of @xmath211 . the absolute bias of galaxies at @xmath0 measured here is higher than at @xmath103 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , where galaxies near @xmath223 are found to have a bias of @xmath224 .
to compare our results on the color dependence of galaxy clustering with the deep2 environment results of @xcite , we show in figure [ relbias ] the relative bias of each color sample used here , relative to the main blue sample , measured on scales @xmath225 .
the main red and blue samples are shown as triangles , while the finer color bins are shown as diamonds .
crosses with dotted error bars are the relative galaxy overdensity as measured by the environment statistic @xmath226 ( described in @xcite ) for galaxies with @xmath227 and @xmath59 , as a function of color , again normalized to the main blue sample .
note that this redshift and magnitude range are not exactly the same as those used in @xcite , therefore our figure [ relbias ] does not look identical to figure 5 in @xcite .
the overall agreement is very good between the relative overdensity and the relative bias from the clustering measures presented here ; these two statistics should be similar but not identical .
the overdensity estimate , @xmath226 , is not measured at a fixed scale but is generally in the range @xmath228 ; we compare to the relative bias from clustering averaged on scales @xmath225 to minimize systematics from scale dependence .
the environment overdensity statistic is in some ways a scale - averaged clustering measure , so the agreement between the two should be good
. however , the environment statistic uses _ all _ neighboring galaxies to define the local overdensity , not just those with the same color or magnitude as the galaxy for which one is measuring the environment .
this is different from the clustering statistic we use , where only galaxies within a given color and magnitude range are used for the auto - correlation function . using all galaxies as a density tracer could potentially cause the environment statistic to be more sensitive to group membership and corresponding small - scale overdensities for populations which tend to be central galaxies instead of satellite galaxies ( i.e. , if there is only one galaxy of that type in an overdensity ) .
as seen in figure [ relbias ] , the agreement between the two statistics is good but not exact . in particular , the clustering results show a continued trend within the blue population of the bluest galaxies being the least clustered , which is not clearly reflected in the environment measure , though the discrepancy is at the 1@xmath141 level ; the broader environment samples in @xcite do show such a downturn .
our clustering - based results on the color - density relation at @xmath0 are very similar to local environment findings in sdss .
@xcite find that sdss galaxies also show a color - density relation within the blue population , but not within the red sequence .
they also find that locally the overdensity of blue galaxies shows little luminosity dependence , while there is strong luminosity dependence within the red galaxy population .
unfortunately , the deep2 volume is not large enough to provide a fair sample of the brightest , rarest red galaxies at @xmath0 , and the error bars on our results for @xmath173 as a function of luminosity within the red sequence are 5 - 15% , which may be large enough to mask a modest luminosity dependence for the range of luminosities we probe .
the general conclusion is that there has not been qualitative evolution in the color - luminosity - overdensity space since @xmath0 . on small scales ( @xmath229 )
random motions within galaxy groups and clusters lead to elongations of the contours along the line of sight , the so - called `` fingers of god '' .
these motions can only be seen in spectroscopic redshift surveys , and at @xmath0 only the deep2 and vvds surveys can currently detect these distortions in .
the virialized motion of the galaxies can be measured using the pairwise velocity dispersion , @xmath230 , which is estimated by modeling as a convolution of the real - space correlation function , , with a broadening velocity function .
the relation between and @xmath119 is usually assumed to be @xmath231dv_{12 } , \label{sigmamodel}\ ] ] where @xmath232 is the distribution function of the relative velocity difference between galaxy pairs along the line of sight ( e.g. , * ? ? ?
this velocity difference is defined as @xmath233 where @xmath234 is the mean radial pairwise velocity of galaxies at separation @xmath69 and @xmath93 is the real - space separation along the line of sight .
here @xmath93 has units of instead of ( this conversion includes a factor of @xmath235 , with @xmath40 , and a factor of @xmath236 ) , such that the same physical velocity at @xmath62 has less of an effect on than at @xmath161 .
to measure the small - scale velocity dispersion on scales @xmath237 , we follow @xcite and in equation [ sigmamodel ] we use @xmath238\ ] ] in both real and redshift space , for values of @xmath239 20 @xmath240 mpc .
we then normalize @xmath241 so that our subsequent fitting will be sensitive to the overall shape of @xmath241 but insensitive to the amplitude .
an exponential form is usually adopted for @xmath232 ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) : @xmath242 where @xmath243 is the mean and @xmath230 is the dispersion of the pairwise peculiar velocities .
we assume an infall model based on the similarity solution of the pair conservation equation @xcite , @xmath244 where @xmath245 and @xmath93 is the radial separation in the real space .
the results are relatively insensitive to the assumed model ; @xmath230 is @xmath13920 - 40% lower if @xmath246 is assumed to be negligible .
to estimate @xmath230 , is modelled using the power - law fits for and @xmath3 for each sample , and we minimize @xmath109 between the observed @xmath241 and modelled @xmath241 for a range of @xmath230 values .
the results are given table 2 , where errors are derived from jacknife samples .
red galaxies are found to have @xmath247 , while blue galaxies have a significantly lower value of @xmath248 .
the error bars are too large to detect significant differences as a function of luminosity within the red or blue galaxy populations or as a function of color for the finer color bins .
green galaxies have a large value , @xmath249 , consistent with red galaxies and 2@xmath141 larger than blue galaxies .
these numbers are similar to , but somewhat smaller than , values found at @xmath161 . using the same infall model , @xcite measure @xmath250 for red galaxies in sdss and @xmath251 for blue galaxies at @xmath161 .
@xcite also find for sdss galaxies that @xmath252 for red galaxies and @xmath253 for blue galaxies .
@xcite measure similar values for galaxies in the 2df redshift survey : @xmath254 for passive galaxies ( defined by spectral type ) and @xmath255 for active , star - forming galaxies .
our finding that red galaxies have stronger fingers of god at @xmath0 than blue galaxies again reflects the result that , on average , red galaxies reside in more overdense regions , such as galaxy groups , than blue galaxies .
peculiar velocities also affect on large scales , where coherent infall of galaxies onto forming structures flattens .
a standard method for quantifying these large - scale redshift - space distortions is to measure the ratio of the quadrupole to monopole moments of the two - point correlation function @xcite .
the two - dimensional correlation function can be decomposed into a sum of legendre polynomials , @xmath256 where @xmath257 is the @xmath258 legendre polynomial and @xmath259 is the cosine of the angle between the line of sight and the redshift separation vector , @xmath84 .
the multipole moments are defined as @xmath260 following @xcite , we define @xmath261 this ratio depends on @xmath262 , where @xmath104 is the linear galaxy bias @xcite , as @xmath263 figure [ qratio ] shows @xmath264 for the main blue and red galaxy samples .
@xmath264 is positive on small scales where the fingers of god are strong and negative on large scales where coherent infall dominates .
the dotted lines indicate @xmath265 predicted on linear scales ( from equation 22 ) , assuming @xmath102 at the mean redshift of each sample ( for the blue sample @xmath266 and for the red sample @xmath267 , corresponding to @xmath268 ) and using the inferred linear galaxy bias for each sample , @xmath269 for blue galaxies and @xmath270 for red galaxies . using @xmath271 increases the predicted @xmath265 values by @xmath1395% . for the blue sample @xmath272 and for the red sample @xmath273 .
both galaxy samples have negative @xmath265 values , on scales @xmath274 for blue galaxies and @xmath275 for red galaxies , indicating coherent infall , though the significance is lower for red galaxies . on the largest scales @xmath265 increases ; however , the error bars increase on these scales as well , which approach the transverse scale of the survey and so may be susceptible to systematic effects . on scales
@xmath276 the blue sample is consistent with the predicted constant @xmath265 value within the noise .
the red sample does not reach the predicted negative value on the largest scales ; however , given the non - negligible pairwise velocity dispersion we do not expect this model to be perfect , as even on scales of 15 , @xmath264 is sensitive to @xmath230 and the form of @xmath232 .
this is especially true for the red galaxy sample , where @xmath230 is greater .
the fact that infall is seen indicates that the clustering of galaxies is dynamic and growing , as expected .
we compare our clustering results for blue and red galaxies with the recent semi - analytic galaxy formation model of @xcite applied to the millennium run n - body dark matter simulation @xcite .
the millennium run follows the dynamical evolution of @xmath277 dark matter particles in a periodic box of side - length @xmath278mpc with a mass resolution per particle of @xmath279 .
cosmological parameters consistent with those measured from the first year wmap data @xcite were used .
the galaxy formation model follows the growth of approximately 25 million galaxies from their birth to the present day and has been tuned to provide a good match to many observed properties of local galaxies , primarily the luminosity function and tully - fisher relation . within this model
the behavior of the observed global galaxy two - point correlation function is well reproduced at @xmath103 , however when split by galaxy color , @xmath103 red galaxies are somewhat too clustered in the model while blue galaxies are under - clustered @xcite .
ours is the first test of the clustering of millennium run galaxies as a function of color at higher redshift . using an ensemble of 15 light - cones ,
each of which matches the geometry of one of the four deep2 survey fields ( see * ? ? ?
* for further details ) we measure the clustering of blue and red galaxies with the same @xmath36 and @xmath280 ranges as our main deep2 color samples .
we define galaxies to be blue or red in the model using the same color cut as used for deep2 galaxies ( equation 1 ) .
if we instead use the minimum of the observed color bimodality in the model galaxies ( which is roughly 0.1 mags blueward of the location for the deep2 galaxies ) our qualitative results and conclusions remain unchanged .
details of each sample used in the mock catalogs are given in table 2 .
the number densities for red and blue galaxies in the model match the observed number densities quite well .
the resulting contours for the semi - analytic model galaxies are shown in figure [ mr_xisp ] .
red model galaxies are more clustered and have stronger fingers of god than the blue model galaxies .
a direct comparison between the red model galaxy and the equivalent deep2 result ( figure [ xisp_all ] ) indicates that a ) the redshift - space distortions are generally similar between the model and data , including both the fingers of god and the coherent infall on large scales , and that b ) the overall clustering amplitude is larger in the model ; e.g. compare the thick @xmath132 lines .
hence , the discrepancy seen at @xmath103 appears to extend back to earlier epochs .
the blue model galaxy shows differences with the data as well .
the left panel of figure [ mr_xisp ] shows coherent infall on large scales but no small - scale redshift - space distortions , unlike the deep2 data .
this is reflected in the low value of @xmath230 found for the model blue galaxies ( see table 2 ) .
the discrepancies seen in figure [ mr_xisp ] are highly significant , as seen in the projected correlation functions shown in figure [ mr_compare ] , where the observed deep2 results are shown for comparison . on scales @xmath2 the blue cloud semi - analytic model correlation function is flat and does not continue to rise as in the data .
in addition , in the quasi - linear regime , @xmath106 , the blue model galaxies are less clustered than the data by @xmath281% .
in contrast , the red galaxies in the model are significantly more clustered than the data , @xmath282% over most scales , as also seen at low redshift .
power - law fits are given in table 2 , where slightly larger scales are used for the model fits ( @xmath283 ) than for the data , as the model blue galaxy is not a power law on smaller scales .
finally , the red - blue cross - correlation function for the model galaxies is shown in figure [ mr_cross ] ( solid line ) along with the cross - correlation function in the deep2 data ( thin dashed line ) and the geometric mean of the red and blue auto - correlation functions in the model ( dot - dash line ) .
the cross - correlation function in the model is clearly lower on small scales ( @xmath155 ) than either the geometric mean or the deep2 data .
this reflects the lack of blue galaxies around red galaxies in the model .
interestingly , a hint of this same flattening of for blue galaxies in this model can be seen at @xmath103 in figure 5 of @xcite ; however they do not plot below scales @xmath284 where for us the effect is most pronounced .
it appears that the discrepancies shown in figures [ mr_compare ] and [ mr_cross ] are due to the incorrect modeling of the physics that determines the colors of satellite galaxies .
if too few model satellite galaxies are blue compared to the data at @xmath0 , then both the model blue auto - correlation function and red - blue cross - correlation function will have lower amplitudes on small scales than seen in the deep2 data , inside the ` one - halo ' regime ( @xmath155 ) .
similarly , an overabundance of red satellites on such scales should also be present , as is seen in figure [ mr_compare ] .
a similar effect is detected at low redshift when comparing the fraction of blue sdss galaxies in groups with these millenium run semi - analytic models .
@xcite find that the blue fraction of galaxies in groups is much higher in the sdss data than in the model , which predicts at @xmath161 that almost all satellite galaxies ( @xmath13985% ) are red .
our results provide a valuable way to test some of the physical assumptions made by the semi - analytic model to better understand the processes that may actually be occurring and their relative importance .
we identify two overly simplified aspects of the model satellite evolution that contribute to the clustering discrepancies shown in figure [ mr_compare ] .
the first is the way in which hot gas is stripped from halos ( @xmath285 ) when the halo falls into the potential of a more massive system . as discussed by @xcite ,
the common assumption in most semi - analytic models is one of _ extreme strangulation _
, where all of the hot gas is instantly stripped and added to that of the larger halo upon accretion .
this simplifying assumption is important for galaxy colors because the dynamical time for a sub - halo ( and hence satellite galaxy ) to merge is usually longer than the time it takes the stellar population to fade and redden , which all model satellites will do in the absence of fresh fuel for star formation .
this provides an overly - efficient channel for blue satellites to transform into red satellites .
@xcite point out that the @xcite model does not include a host halo mass dependence for the strangulation efficiency .
given that the ability of a parent halo to strip gas from a subhalo likely depends on the density of the hot gas in the parent halo , a model which includes such a mass - dependent strangulation prescription might provide a better match to the results shown here .
the second aspect of the model satellite evolution that may be discrepant with the real universe is the assumption that local potential wells ( i.e. subhalos inside larger parent halos ) can not accrete gas , meaning that after a satellite galaxy is captured by the larger halo it receives no new baryonic fuel for later star formation from the intragroup or intracluster medium .
this may be generally true in n - body simulations which show that subhalos tend to lose mass through tidal stripping as they spiral inward , not gain mass .
however it is plausible that in the outer regions of a dark matter halo , the local potential of a subhalo may dominate that of its host and attract new baryons from the hot parent halo , baryons that may subsequently cool onto the satellite .
such gas would extend the star formation history of the satellite and hence modify the evoution of its color .
it is also important to note that there are processes that are known to operate in cluster ( and also group ) environments that act to remove cold gas from a satellite galaxy , such as tidal stripping and harassment .
these processes are not included in the @xcite semi - analytic model used here ( most semi - analytic models ignore such additional effects to keep the models relatively simple ) .
their inclusion would typically redden the color of the satellite further with time , worsening an effect that is already too efficient in the current model .
our work indicates that current prescriptions for satellite infall and its consequent effect on star formation may be too simplistic .
however , it also highlights the utility of color - dependent clustering measurements in constraining such models .
in this paper we use volume - limited subsamples of the deep2 galaxy redshift survey data to measure the color and luminosity dependence of galaxy clustering at @xmath0 .
we split the sample into blue , star - forming galaxies and red , nonstar - forming galaxies using the observed color bimodality in @xmath37 .
we further subdivide each sample into luminosity and finer color bins .
* red galaxies are much more strongly clustered than blue galaxies at @xmath0 , with a relative bias that is roughly as high as is found locally ( @xmath286 ) .
this implies that , for relatively bright galaxies , the color - density relation was as strongly in place 8 gyr ago as it is observed to be today .
* the comoving clustering amplitudes that we measure for both blue and red galaxies are only slightly ( 10 - 20% ) lower than what is found locally .
the measured values are somewhat higher than those in previous deep2 and vvds results @xcite , though generally within 2@xmath141 . *
we find no statistically - significant dependence of galaxy clustering on luminosity for either red or blue galaxies , within the luminosity range that we probe .
additionally , within the red sequence there is no dependence of clustering on color , while within the blue cloud there is a strong dependence on color .
the color dependence and lack of luminosity dependence for both red and blue that we find is consistent with previous results at @xmath161 , for the luminosity range probed here .
* the stronger dependence of clustering on luminosity found for all galaxies in the deep2 sample @xcite is likely due in part to the changing fraction of red versus blue galaxies as a function of luminosity , as we do not find as strong of a luminosity dependence here within either the red or blue galaxy populations independently .
* the brightest blue galaxies have a significant rise in their correlation function on small scales ( @xmath287 ) , which likely reflects a lower satellite fraction compared to fainter blue galaxies ; i.e. , bright blue galaxies are more likely to be central galaxies within their parent dark matter halos . *
the correlation between color and clustering within the blue cloud is likely a reflection of a correlation between stellar mass and halo mass .
the redder galaxies in the blue cloud have higher stellar masses ( see figure 8 of * ? ? ?
* ) and are significantly more clustered than the bluer galaxies , which implies that they are in more massive dark matter halos . * red galaxies show strong ` fingers of god ' at @xmath0 , indicating that they lie in virialized overdensities such as groups and clusters .
blue galaxies show smaller ` fingers of god ' .
* both blue and red galaxies show a flattening of on large scales due to coherent infall of galaxies onto structures that are still collapsing ( the kaiser effect ) . * the projected correlation functions of brighter samples near @xmath223 show deviations from a power law on small scales , within the one - halo regime .
the deviations are more pronounced for blue galaxies than for red galaxies . *
the red - blue cross - correlation function is consistent with the geometric mean of the auto - correlation functions of blue and red galaxies separately .
there is a marginal trend on small scales for the cross - correlation to lie below the geometric mean .
such a feature , if real , would indicate a lack of blue galaxies near red galaxies .
this could reflect a suppression of star formation in the centers of overdensities , such as groups , which are dominated by red galaxies at @xmath0 as seen in the group - galaxy cross - correlation function @xcite . * green galaxies in the valley of the observed color bimodality display a similar large - scale clustering strength as red galaxies , but show a small - scale clustering amplitude and infall kinematics akin to blue galaxies .
the green galaxy population does not have strong ` fingers of god ' but does show coherent infall on large scales , and the correlation function has a similar slope to the blue galaxy population .
green galaxies thus appear to be at the edges , but not in the centers , of the same overdensities as red galaxies .
this is consistent with green galaxies being transition objects moving from the blue cloud to the red sequence and supports a picture of quenching happening preferentially on the outskirts of overdensities at @xmath0 on scales @xmath288 .
* we provide a direct test of environment and clustering statistics , comparing the relative bias as a function of color using a local galaxy overdensity estimator with the two - point correlation function results , and find good agreement between the two . * comparing with the semi - analytic galaxy evolution model of @xcite , we find that red model galaxies are significantly more clustered than the deep2 galaxies , and that in the model the clustering strength of blue galaxies is too low , especially on small scales ( @xmath2 ) .
these discrepancies are likely due to galaxies being quenched too efficiently in the model ; i.e. , star formation is shut down in satellite galaxies too quickly .
* these results all rely on precise , robust , spectroscopic redshifts .
the deep2 rms redshift errors ( determined from repeated observations ) are @xmath34 km s@xmath35 , which is unprecedented at @xmath0 and makes the analyses presented here possible . from the observed bias of the main red and blue galaxy samples we can infer the minimum dark matter mass for halos that host these galaxies . under the assumption that all galaxies are central galaxies and not satellites ( i.e. , one galaxy per dark matter halo ) , the large scale bias of the galaxy population can be matched to the bias of dark matter halos using the formulae from e.g. , mo & white ( 1996 ) or sheth , mo & tormen ( 2001 ) .
however , we know that not all of the deep2 galaxies are central galaxies @xcite , and to obtain a more precise estimate of the minimum dark matter halo mass we use the results presented in the appendix of @xcite , which take into account the inferred satellite fraction from hod modeling of the deep2 luminosity - dependent clustering results @xcite . at @xmath162 , the observed bias and number densities of the main red and blue galaxy samples ( with @xmath289 ) imply minimum dark matter halo masses of @xmath290 for red galaxies and @xmath291 for blue galaxies .
however , the halo occupation of red and blue galaxies is likely complicated enough to warrant a full hod model , which we defer to a future paper ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* for modeling of local red and blue samples ) .
the different slopes measured for the clustering of red and blue galaxies would pose a potential difficulty in terms of understanding how red and blue galaxies populate dark matter halos , if the slopes were different on large scales ( @xmath292 ) . on these scales the halo bias is linear , in that there is an amplitude offset between for halos of different mass , but no difference in the slope ( e.g. , * ? ? ?
* ; * ? ? ?
we do not have a large enough survey here to claim a significant difference in the slope on scales @xmath292 , though the upper left panel of figure [ wprp_all ] does suggest such a difference .
it has been previously established that the number density of galaxies on the red sequence has increased since @xmath62 @xcite .
it is thought that these galaxies likely migrated from the blue cloud to the red sequence as their star formation was quenched . using the observed clustering of red and blue galaxies at @xmath0
, one can predict what their clustering strength should be at @xmath293 , for a given cosmology and assuming no merging of galaxies , by using the continuity equation and the expected growth of perturbations @xcite .
this leads to larger predicted clustering amplitudes than is observed for local @xmath223 red and blue galaxies ; the main red and blue samples used here ( both of which have @xmath179 at @xmath162 ) would have clustering scale - lengths of @xmath294 and @xmath295 , respectively , at @xmath103 .
the fact that local clustering measurements at @xmath223 are lower than these predictions , especially for blue galaxies , lends credence to the idea that blue galaxies have migrated to the red sequence and less clustered blue galaxies have joined the blue cloud since @xmath62 .
the color and luminosity dependence of clustering at @xmath62 can illuminate _ which _ blue galaxies likely become red between @xmath62 and @xmath103 .
the fact that the relative bias between red and blue galaxies is as high at @xmath0 as @xmath293 indicates that blue galaxies have to be turning red _ in a density - dependent way _ , in that the more clustered of the blue galaxies must become red .
the bias for any galaxy population should tend towards unity with time @xcite and linear growth theory would predict that the relative bias of our red and blue samples should decrease by 8% to @xmath161 .
observationally , however , it appears that the relative bias between red and blue galaxies has either remained constant since @xmath0 or increased .
this therefore implies that the most clustered blue galaxies must be turning red to boost the clustering of red galaxies relative to blue galaxies and to counteract the effect of the relative bias decreasing .
it is the most clustered of the blue galaxies at @xmath62 that are likely to be as clustered as red galaxies by @xmath103 , and our results here indicate that the reddest of the blue galaxies and the green galaxies are the most likely candidates for having their star formation quenched by @xmath103 , based on their clustering properties . the luminosity dependence of clustering at @xmath0 can be explained as being almost entirely due to the relationship between luminosity and halo mass , as shown by @xcite who use a simple one - to - one relation between halo mass and light to obtain the same luminosity- and scale - dependence of clustering in an n - body simulation as we find in deep2 . the tight relation between halo mass and galaxy light
is confirmed by @xcite , who fit hod models to the deep2 luminosity dependent clustering results .
however , the color dependence of clustering seen here can not entirely be halo - mass dependent , as the clustering as a function of color does not match the results of @xcite , where the galaxies have been ranked by luminosity , which is equivalent to halo mass ; i.e. , the trends seen with color in the data do not match the trends found with luminosity or halo mass in the model .
this implies that there is a wider range in the @xmath296 ratio of galaxies at a given color than at a given luminosity . to constrain galaxy evolution physics it may therefore be more informative to study samples divided by color rather than luminosity , as more galaxy physics
is apparently needed to interpret the color dependence of galaxy clustering .
the lack of a stronger luminosity dependence for the separate red and blue galaxy samples shown here makes the results of @xcite more surprising , as it is not clear why @xmath36 luminosity for all galaxies should be so tightly correlated with halo mass .
it may be that the relative mix of red and blue galaxies as a function of luminosity helps to create a stronger relation with luminosity for all galaxies and is more tightly correlated with halo mass than the luminosity itself .
as noted in previous papers on the color - density relation in the deep2 sample @xcite , the observed color - density relation at @xmath0 can not be the result of cluster - specific physics as only a few percent of deep2 galaxies are in massive clusters .
the color dependence of clustering must be dominated by either processes that operate in a group environment or intrinsic processes that depend on the age , stellar mass or halo mass of a galaxy .
processes that occur when a galaxy is accreted into a larger halo , becoming a satellite galaxy , may _ contribute _ to the observed color - density relation but do not _ dominate _ it , as most deep2 galaxies are not satellite galaxies @xcite .
what star - formation quenching processes may be at work in groups ?
there are more galaxy mergers in groups than in clusters , such that mergers may play an important role in shutting off star formation in group galaxies .
there is less hot ambient gas in groups than in clusters ; it is not clear whether the density of the intragroup gas is high enough to strip gas from infalling satellite galaxies , except near the core of the group .
apparently not all satellite galaxies are red , as is seen in the clustering results ; this implies that there is a non - negligible timescale for quenching star formation in satellite galaxies .
deep2 galaxies in groups at @xmath0 have a clustering scale - length of @xmath297 and a slope of @xmath298 @xcite , similar to red galaxies as measured here .
galaxies in the field at @xmath0 have a slope of @xmath299 , similar to blue galaxies , but with a lower correlation length , @xmath300 .
the cross - correlation of red galaxies and groups is higher than that of blue galaxies and groups @xcite , which also indicates that red galaxies are more likely to be in groups . however , the picture is not as simple as all red galaxies being in groups and blue galaxies being in the field : blue galaxies have some fingers of god on small scales , indicating that at least some are in virialized overdensities , and the brighter blue galaxies show a rise in the correlation function on small scales , which is probably due to a lower satellite fraction for the brighter blue samples .
these galaxies are likely the central galaxies in groups at @xmath62 , and may migrate to the red sequence by @xmath293 .
the scale at which the one - halo term rise occurs for these samples is small , @xmath301 ( as seen in figure [ wprp_lum_powerlaw ] ) , and indicates again that these galaxies are likely to be in the centers of the groups they reside in .
additionally , @xmath302% of deep2 galaxies in underdense regions are red @xcite .
the green galaxy population , which lies near the minimum in the observed color bimodality , appears to be , at least in part , a distinct population and is not a simple mix of red and blue galaxies .
though the green population can include the tail of each of the blue and red distributions , it also contains transitional objects , likely moving from the blue cloud to the red sequence .
the green galaxies have different clustering characteristics than either the red or blue galaxies as a whole . on large scales ( @xmath131 )
the green galaxies are as clustered as red galaxies of the same luminosity , implying that they reside in the same relatively massive halos as red galaxies . on small scales
, however , the clustering amplitude is similar to that of blue galaxies , which likely reflects a lower radial concentration for green galaxies compared to red galaxies .
the general picture that emerges is one in which star formation quenching is occurring primarily at the edges of overdensities as galaxies fall in , turning blue galaxies green around the edges of groups .
we have identified that the scale at which this transition from red to blue occurs at @xmath0 is 0.5 @xmath303 2 .
this scale is just larger than the transition scale between the one - halo and two - halo terms at @xmath0 , which is shown by @xcite ( see their fig .
1 ) to be @xmath11 for all galaxies and by @xcite ( see their fig . 8) to be @xmath304 for galaxies in groups .
this transition scale for green galaxies is therefore similar to or just larger than the largest halos , which indicates that it is occurring at @xmath305 virial radii , on the edges of large halos .
this conclusion lends support to the idea that the mechanism which shuts off star formation in these galaxies is not happening at the centers of massive halos , but on the outskirts .
the comparison to semi - analytic galaxy formation models demonstrates vividly the utility of using correlation statistics to test theoretical models , and the need to test them at intermediate redshift , not just locally .
it is also likely that currently popular models that have a simple halo threshold mass for quenching star formation in galaxies will suffer similar problems as the model shown here .
we urge theorists to use these data to constrain their models .
we thank charlie conroy , daniel eisenstein , jeremy tinker , frank van den bosch , christopher willmer and zheng zheng for useful discussions and comments on this paper .
alc and np are supported by nasa through hubble fellowship grants hf-01182.01-a and hst - hf-01200 , respectively , awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa , under contract nas 5 - 26555 .
the deimos spectrograph was funded by a grant from cara ( keck observatory ) , an nsf facilities and infrastructure grant ( ast92 - 2540 ) , the center for particle astrophysics and by gifts from sun microsystems and the quantum corporation .
the data presented herein were obtained at the w.m .
keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration .
the observatory was made possible by the generous financial support of the w.m .
keck foundation . the deep2 team and keck observatory
acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community and appreciate the opportunity to conduct observations from this mountain .
llllllll norberg et al .
2002 & 0.1 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & & @xmath310 + madgwick et al .
2003 & 0.1 & @xmath311 & @xmath312 & @xmath313 & @xmath314 & @xmath315 & @xmath169 + budavari et al .
2003 & 0.2 & @xmath1394.5 - 6.0 & @xmath316 & @xmath317 & @xmath318 & & @xmath319 + zehavi et al . 2004 & 0.1 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath213 & @xmath324 + coil et al .
2004 & 0.8 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & + meneux et al .
2006 & 0.8 & @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & + [ published ] lrcccclcccccl red : -21.0 & 792 & @xmath335 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath342 & @xmath343 & @xmath344 + red : -20.5 & 1370 & @xmath345 & @xmath346 & @xmath347 & @xmath348 & @xmath349 & @xmath350 & @xmath351 & @xmath352 & @xmath353 & @xmath354 + red : -20.0 ( main ) & 1474 & @xmath355 & @xmath346 & @xmath356 & @xmath357 & @xmath358 & @xmath359 & @xmath360 & @xmath361 & @xmath362 & @xmath363 + red : -19.5 & 1032 & @xmath364 & @xmath365 & @xmath366 & @xmath367 & @xmath368 & @xmath369 & @xmath370 & @xmath371 & @xmath372 & @xmath373 + blue : -21.0 & 1424 & @xmath374 & @xmath375 & @xmath337 & @xmath376 & @xmath377 & @xmath378 & @xmath379 & @xmath380 & @xmath381 & @xmath382 + blue : -20.5 & 3747 & @xmath355 & @xmath369 & @xmath347 & @xmath383 & @xmath377 & @xmath384 & @xmath385 & @xmath386 & @xmath387 & @xmath388 + blue : -20.0 ( main ) & 4808 & @xmath389 & @xmath167 & @xmath356 & @xmath390 & @xmath391 & @xmath392 & @xmath393 & @xmath394 & @xmath395 & @xmath396 + blue : -19.5 & 4018 & @xmath397 & @xmath398 & @xmath366 & @xmath399 & @xmath400 & @xmath375 & @xmath401 & @xmath402 & @xmath403 & @xmath404 + red : redder & 731 & @xmath405 & @xmath406 & @xmath356 & @xmath407 & @xmath358 & @xmath375 & @xmath408 & @xmath409 & @xmath410 & @xmath411 + red : bluer & 743 & @xmath412 & @xmath413 & @xmath356 & @xmath414 & @xmath358 & @xmath359 & @xmath415 & @xmath416 & @xmath417 & @xmath418 + green & 997 & @xmath419 & @xmath378 & @xmath356 & @xmath420 & @xmath421 & @xmath422 & @xmath423 & @xmath424 & @xmath425 & @xmath426 + blue : redder & 1407 & @xmath427 & @xmath350 & @xmath356 & @xmath420 & @xmath421 & @xmath350 & @xmath428 & @xmath429 & @xmath430 & @xmath431 + blue : middle & 1204 & @xmath432 & @xmath433 & @xmath356 & @xmath434 & @xmath421 & @xmath422 & @xmath435 & @xmath436 & @xmath437 & @xmath438 + blue : bluer & 1214 & @xmath432 & @xmath439 & @xmath356 & @xmath440 & @xmath421 & @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath445 + mr+sam red : & 7875 & @xmath446 & @xmath447 & @xmath356 & @xmath448 & @xmath421 & @xmath449 & @xmath450 & @xmath451 & @xmath452 & @xmath453 + mr+sam blue : & 23645 & @xmath454 & @xmath455 & @xmath356 & @xmath456 & @xmath421 & @xmath350 & @xmath457 & @xmath458 & @xmath459 & @xmath460 + [ samples ] | we present measurements of the color and luminosity dependence of galaxy clustering at @xmath0 in the deep2 galaxy redshift survey . using volume - limited subsamples in bins of both color and luminosity
, we find that : 1 ) the clustering dependence is much stronger with color than with luminosity and is as strong with color at @xmath0 as is found locally .
we find no dependence of the clustering amplitude on color for galaxies on the red sequence , but a significant dependence on color for galaxies within the blue cloud .
2 ) for galaxies in the range @xmath1 , a stronger large - scale luminosity dependence is seen for all galaxies than for red and blue galaxies separately .
the small - scale clustering amplitude depends significantly on luminosity for blue galaxies , with brighter samples having a stronger rise on scales @xmath2 .
3 ) redder galaxies exhibit stronger small - scale redshift - space distortions ( `` fingers of god '' ) , and both red and blue populations show large - scale distortions in due to coherent infall .
4 ) while the clustering length , , increases smoothly with galaxy color ( in narrow bins ) , its power - law exponent , @xmath3 , exhibits a sharp jump from the blue cloud to the red sequence .
the intermediate color ` green ' galaxy population likely includes transitional galaxies moving from the blue cloud to the red sequence ; on large scales green galaxies are as clustered as red galaxies but show infall kinematics and a small - scale correlation slope akin to the blue galaxy population .
5 ) we compare our results to a semi - analytic galaxy formation model applied to the millenium run simulation .
differences between the data and the model suggest that in the model star formation is shut down too efficiently in satellite galaxies . |
there are sound theoretical and experimental reasons for studying flavor changing neutral current ( fcnc ) processes .
such transitions , being forbidden at tree level , provide stringent tests of the standard model ( sm ) at the loop level . besides , fcncs form a natural arena for discovering indirect effects of possible tev scale extensions of the sm such as supersymmetry . among all the fcnc phenomena , the rare decays of the @xmath8 mesons are particularly important as many of the nonperturbative effects are small and under control .
in addition to having already determined the branching ratio of @xmath11 @xcite and the cp asymmetry of @xmath12 @xcite , experimental activity in @xmath8 physics , has begun to probe fcnc phenomena in semileptonic @xmath8 decays @xcite .
therefore , with increasing data and statistics , these experiments are expected to give precise measurements on long and short distance effects in semileptonic decays , @xmath13 @xmath14 ( @xmath15 ) and @xmath16 ( @xmath17 ) . the key physical quantities that can be measured are the branching ratios , cp asymmetries and several lepton asymmetries . in searching for the physics beyond the sm
it is often necessary to deal with quantities that differ significantly from their sm counterparts .
this is because there are large uncertainties coming from the hadronic form factors making it hard to disentangle new physics effects from those of the hadronic dynamics .
for this reason , the pseudoscalar channel @xmath14 provides a unique opportunity , since the forward backward asymmetry @xmath18 in this channel
is extremely small in the sm ( due to a suppression of order @xmath19 ) , and this remains true in any of its extensions unless a scalar scalar type four fermion operator such as @xmath20 provides a significant contribution to the decay amplitude .
clearly , such operator structures can arise only from the exchange of a scalar between the quark and lepton lines with flavor violating couplings to the quarks .
for instance , by extending the sm higgs sector to two su(2 ) doublets , operator structures of the form ( [ oper ] ) can be generated @xcite excluding the possibility of @xmath21 @xmath22 tree level fcncs .
although the coefficient @xmath23 in eq.([oper ] ) is still proportional to the lepton mass , it can receive an enhancement when the ratio of the two higgs vacuum expectation values , @xmath24 is large .
supersymmetry ( susy ) , is one of the most favoured extensions of the sm which stabilizes the scalar sector against ultraviolet divergences , and naturally avoids the dangerous tree level fcnc couplings by coupling the higgs doublet @xmath25 ( @xmath26 ) to up
type quarks ( down type quarks and charged leptons ) .
the soft breaking of susy at the weak scale generates ( @xmath27 ) a variety of new sources for tree level flavor violation depending on the structure of the soft terms , and ( @xmath28 ) radiatively generates various fcnc couplings even if the flavor violation is restricted to the ckm matrix . the first effect , which can not be determined theoretically , is strongly constrained by the fcnc data @xcite , and therefore , as a predictive case , it is convenient to restrict all flavor violating transitions to the charged current interactions where they proceed via the known ckm angles .
this is indeed the case in various susy breaking schemes where hidden sector breaking is transmitted to the observable sector via flavor
blind interactions @xmath13 gauge mediated and minimal gravity mediated scenarios .
this minimal flavor violation scheme adopted here is well motivated by minimal supergravity in which all scalars receive a common soft mass , @xmath29 , at the unification scale .
the common origin for scalar masses is one of the parameter restrictions which define the constrained version of the supersymmetric standard model ( cmssm ) .
the low energy sparticle spectrum in the cmssm is specified entirely by four parameters , and one sign .
in addition to @xmath29 and @xmath30 , the remaining mass parameters are the gaugino masses and supersymmetry breaking trilinear mass terms .
these too , are assumed to have common values , @xmath31 and @xmath32 , at the unification scale . in principle , there are two additional parameters , the higgs mixing mass , @xmath4 , and the supersymmetry breaking bilinear mass term , @xmath8 , but since it is common to choose @xmath33 as a free parameter and since we fix the sum of the squares of the two higgs vevs with @xmath34 , these two parameters are fixed by the requirements of low energy electroweak symmetry breaking .
one is left simply with a sign ambiguity for @xmath4 .
therefore , the parameters which define a cmssm model are @xmath35 and @xmath36 .
although flavor violation is restricted to the ckm matrix , radiative effects still generate fcnc transitions among which those that are enhanced at large values of @xmath3 are particularly important as the lep era ended with a clear preference to large values of @xmath3 @xcite .
indeed , it is known that there are large @xmath3enhanced threshold corrections to the ckm entries @xcite allowing for higgs mediated fcnc transitions @xcite .
for instance , the holomorphic mass term for down type quarks @xmath37 acquires a non - holomorphic correction @xmath38 where the latter term proportional to @xmath39 , which is not necessarily small at large @xmath3 . in what follows
, we will compute the forward backward asymmetry of @xmath40
decays in the mssm . after deriving the scalar exchange amplitudes ( [ oper ] )
, we discuss several theoretical and experimental issues and then identify the regions of susy parameter space for which the asymmetry is enhanced .
we compare our results with existing experimental and cosmological constraints .
in general , the semileptonic decays @xmath40 proceed via the quark transitions @xmath41 .
the decay amplitude has the form @xmath42\end{aligned}\ ] ] where @xmath43 with @xmath44 being the dilepton invariant mass .
the wilson coefficients , @xmath45 , and @xmath46 have been computed to leading order in @xcite .
higher order @xmath47 corrections , which are available for small @xmath48 in the sm @xcite , will not be considered .
the coefficients @xmath49 and @xmath50 will be discussed below .
the kinematical range for the normalized dilepton invariant mass in terms of the lepton and pseudo scalar masses is @xmath51 which includes the vector charmonium resonances @xmath52 whose effects are included in the @xmath53 .
moreover , the four fermion operators for the light quarks develop nonvanishing matrix elements , and these are also included in @xmath53 . at higher orders in @xmath54 , these effects contribute to @xmath55 as well @xcite . the electromagnetic dipole coefficient @xmath55 is contributed by graphs with the w boson , charged higgs , and chargino penguins .
the chargino contribution increases linearly with @xmath3 at leading order @xcite , and the inclusion of susy threshold corrections strengthens this dependence @xcite .
this coefficient is directly constrained by the @xmath56 decay rate , and the experimental bounds can be satisfied with a relatively light charged higgs at very large values of @xmath3 . on the other hand , the coefficient of the vector
vector operator @xmath53 is generated by box diagrams , and carries a long distance piece coming from the matrix elements of the light quark operators as well as the intermediate charmonium states @xcite . finally , the coefficient of the vector
pseudovector operator @xmath57 is generated by box graphs and is scale independent .
both coefficients @xmath53 and @xmath57 are less sensitive to @xmath3 than is @xmath55 . within the sm
, these coefficients typically have the values @xmath58 , @xmath59 ( excluding its long distance part ) , and @xmath60 @xcite which , however , are allowed to vary considerably within the existing bounds @xcite . the inclusion of susy contributions , for instance , implies large variations in @xmath55 ( even changing its sign ) , and typically a @xmath61 variation in @xmath53 and @xmath57 @xcite .
the scalar
scalar operators in the decay amplitude are generically induced by the exchange of the higgs scalars and suffer invariably from the @xmath62 suppression .
therefore , these operators are completely negligible in the sm . however , in the mssm , this suppression is overcome by large @xmath3 effects where the charged higgs top diagram is proportional to @xmath63 , and the chargino
stop diagram is @xmath64 . in more explicit terms , @xmath65 , and @xmath66\end{aligned}\ ] ] where @xmath67 is the sign of the @xmath4 parameter
, @xmath68 is the lighter chargino , @xmath69 , @xmath70 , and the parameters @xmath71 and @xmath72 , which are typically @xmath73 , are defined in @xcite .
finally , @xmath74 is the top quark yukawa coupling , @xmath75 is the light stop mass , and @xmath76 is the low energy value of the susy breaking top - yukawa trilinear mass term obtained from @xmath32 by the running of the rges
. clearly , the charged higgs contribution , which is the dominant one in two doublet models @xcite , is subleading compared to the chargino contribution .
the sign of @xmath23 depends explicitly on @xmath77 .
therefore , the forward backward asymmetry in @xmath0 decays depends strongly on the sign of the @xmath4 parameter . from the experimental point of view , it is useful to analyze the normalized forward backward asymmetry defined as @xmath78 where @xmath79 , @xmath80 being the angle between the momenta of @xmath81 and @xmath82 .
a direct calculation gives the explicit expression @xmath83 a(s)}{\sigma(s)}\end{aligned}\ ] ] where @xmath84 + t_{\ell } t_{p } a_{10}(s)^{2}\nonumber\\ & + & s t_{\ell}\left[(b_{10}(s)-a(s))^{2}+v(s)^{2 } a(s)^{2}\right]\nonumber\\&+ & 2 t_{\ell } ( 1-t_{p}/4-s ) [ b_{10}(s)-a(s ) ] a_{10}(s)\;.\end{aligned}\ ] ] here @xmath85 , @xmath86 , @xmath87 , @xmath88 and @xmath89\end{aligned}\ ] ] where @xmath90 .
the form factors @xmath91 , @xmath92 and @xmath93 are not measured at present and one has to rely on theoretical predictions . in what follows we use the results of the calculation @xcite of these form factors from qcd sum rules for both @xmath94 and @xmath95 transitions . in general ,
the hadronic form factors are uncertain by @xmath96 , and this translates into an uncertainty of approximately @xmath97 in the branching ratio . especially for low @xmath48 , below the charmonium resonances ,
the theoretical prediction for the branching ratio contains large uncertainties @xcite . therefore , theoretically the large dilepton mass region is more tractable . on top of the form factor uncertainties , there are further problems in treating the contributions of the charmonium resonances ( embedded in the wilson coefficient @xmath53 ) . for instance , the recent belle experiments @xcite , subtract such resonance contributions by vetoing the range @xmath98 .
then the experimental bound on the branching ratio turns out to be @xmath99 which we will take into account in making the numerical estimates below .
it can be also noted that the same decay mode has not been observed by babar : @xmath100 @xcite .
in addition , the vector kaon final states have not been observed yet : @xmath101 @xcite .
one notes that , the asymmetry is large in regions of the parameter space where the branching ratio is depleted , and therefore , the belle lower bound on @xmath102 is an important constraint which can prohibit the asymmetry taking large marginal values . clearly , in the presence of @xmath23 , which can take large values in susy , the would
be experimental constraints on the @xmath103@xmath53 plane are lifted .
furthermore , the pure leptonic decay modes , @xmath104 , depend directly on the wilson coefficients @xmath103 , @xmath23 and @xmath105 . in the sm ,
@xmath106 which is approximately three orders of magnitude below the present bounds @xmath107 @xcite . the susy contributions , especially at large @xmath3 , can enhance the sm prediction typically by an order of magnitude , and the bounds can even be violated in certain corners of the parameter space @xcite . in what follows the constraints from @xmath108 as well as the muon @xmath10 ( as they are directly correlated @xcite )
will be taken into account .
we will refer to the constraints from @xmath109 and @xmath104 collectively as @xmath8factory constraints .
.2 in for the constraints on the susy parameter space to make sense it is necessary to be far from the regions of large hadronic uncertainties , and thus , below we will restrict the range of @xmath48 to lie well above the charmonium resonances and well below the kinematical end point . in fig .
[ fig6 ] , we show the variation of the asymmetry with the normalized dilepton invariant mass , @xmath110 for various values of the susy parameters ( see below for further discussion of these choices ) .
the irregularities in the @xmath48dependence of the asymmetry are similar to those in the @xmath111 decay .
the various bumps and valleys come from the relative sizes of individual terms contributing to @xmath112 .
it should be noted that in the region around the value @xmath113 , the @xmath48 dependence of the asymmetry is rather smooth .
therefore , in forming the constant asymmetry contours in the space of susy parameters we will take @xmath113 ( corresponding to @xmath114 ) . at low values of the asymmetry one should also take into account the final state electromagnetic interactions .
indeed , photon exchange between the lepton and @xmath115 lines is expected to induce an asymmetry @xmath116 implying that only asymmetries @xmath112 larger than @xmath117 can be trusted to follow from susy effects , unless the interplay with the electromagnetic corrections is explicitly taken into account . for large values of asymmetry ,
when the observation of the effect becomes feasible , higher order qcd effects ( not yet calculated ) can in principle modify our results somewhat .
however , it is highly unlikely that these corrections will dramatically reduce the asymmetry discussed here .
one should also note that in the limit of exact @xmath118 flavor symmetry , the asymmetries in @xmath119 and @xmath120 decays must be the same . due to @xmath118 breaking effects , which show up in different parameterizations of the form factors @xmath91 , @xmath92 and @xmath121 for @xmath95 and @xmath94 transitions @xcite
, their asymmetries are expected to differ slightly .
clearly it is the lepton flavor that largely determines the size of the asymmetry , rather than the flavor of the final state pseudoscalar .
finally , before starting to scan the susy parameter space , it is worth while discussing the sensitivity of @xmath112 to some of the parameters .
first of all , as @xmath3 grows , two of the wilson coefficients , @xmath55 and @xmath23 , grow rapidly up to the bounds obtained from rates of the decays @xmath122 and @xmath123 . since @xmath124 > 0 $ ] and @xmath125 , @xmath126 $ ] increases with @xmath3 . however , this increase is much milder than the @xmath127 dependence of @xmath23 causing @xmath128 to take large negative ( positive ) values for @xmath5 ( @xmath6 ) . therefore , large @xmath3 effects influence not only the numerator of eq .
( [ afb ] ) but also the denominator @xmath129 ( proportional to the differential branching fraction ) via the destructive ( constructive ) interference with @xmath130 ( @xmath103 remains negative in susy ) for @xmath5 ( @xmath6 ) .
however , as @xmath3 keeps growing , depending on the rest of the susy parameters , the effect of @xmath23 eventually becomes more important , and the asymmetry falls rapidly due to the enhanced branching ratio . in this sense ,
the regions of enhanced asymmetry depend crucially on the sign of the @xmath4 parameter and the specific value of @xmath3 .
moreover , as the expression of @xmath23 makes clear , there can be sign changes in the asymmetry in certain regions of the parameter space due to the relative sizes of the masses of the lighter chargino and stops .
such effects will also give small asymmetries just like the @xmath6 case .
our work extends a previous analysis of this asymmetry @xcite by including the large gluino exchange effects ( contained in the quantity @xmath131 in eq .
( [ cmb ] ) ) and the explicit dependence on the sign of the @xmath4 parameter .
in addition , we go beyond the work in @xcite as well as in the preceding work @xcite by resumming the higher order @xmath3 terms which increases the validity of the analysis at large values of @xmath3 @xcite .
we note that a computation of the large @xmath3 effects can be carried out in the gaugeless limit @xcite which eliminates some of the diagrams considered in @xcite . in the numerical analysis
below , we will analyze the forward
backward asymmetry of the decays @xmath14 by taking into account the above mentioned constraints from @xmath8 factories as well as other collider and cosmological constraints .
we will be searching for those regions of the susy parameter space in which the asymmetry is enhanced .
in particular , we will be particularly interested in the sensitivity of the asymmetry to @xmath3 , the sign of the @xmath4 parameter , as well as the common scalar mass @xmath29 and the gaugino mass @xmath31 .
in our analysis , we include several accelerator as well as cosmological constraints . from the chargino searches at lep @xcite , we apply the kinematical limit @xmath132 104 gev .
a more careful consideration of the constraint would lead to an unobservable difference in the figures shown below .
this constraint can be translated into a lower bound on the gaugino mass parameter , @xmath31 and is nearly independent of other susy parameters .
the lep chargino limit is generally overshadowed ( in the cmssm ) by the important constraint provided by the lep lower limit on the higgs mass : @xmath133 114.1 gev @xcite .
this holds in the standard model , for the lightest higgs boson @xmath134 in the general mssm for @xmath135 , and almost always in the cmssm for all @xmath3 .
the higgs limit also imposes important constraints on the cmssm parameters , principally @xmath31 , though in this case , there is a strong dependence on @xmath33 .
the higgs masses are calculated here using feynhiggs @xcite , which is estimated to have a residual uncertainty of a couple of gev in @xmath136 .
we also include the constraint imposed by measurements of @xmath137 @xcite .
these agree with the standard model , and therefore provide bounds on mssm particles , such as the chargino and charged higgs masses , in particular . typically , the @xmath137 constraint is more important for @xmath138 , but it is also relevant for @xmath139 , particularly when @xmath3 is large .
the final experimental constraint we consider is that due to the measurement of the anomalous magnetic moment of the muon .
the bnl e821 @xcite experiment reported a new measurement of @xmath140 which deviates by 1.6 standard deviations from the best standard model prediction ( once the pseudoscalar - meson pole part of the light - by - light scattering contribution @xcite is corrected ) .
although negative values of @xmath4 are no longer entirely excluded @xcite , the 2-@xmath141 limit still excludes much of the @xmath138 parameter space @xcite .
@xmath142 is allowed so long as either ( or both ) @xmath31 and @xmath29 are large .
we also apply the cosmological limit on the relic density of the lightest supersymmetric particle ( lsp ) , @xmath143 , and require that 0.1 < _ h^2 < 0.3 [ ten ] the upper limit is rigorous , and assumes only that the age of the universe exceeds 12 gyr .
it is also consistent with the total matter density @xmath144 , and the hubble expansion rate @xmath145 to within about 10 % ( in units of 100 km / s / mpc ) . on the other hand ,
the lower limit in ( [ ten ] ) is optional , since there could be other important contributions to the overall matter density .
the cosmologically allowed regions in the cmssm have been well studied @xcite .
there are generally large , ` bulk ' regions of parameter space at low to moderate values of @xmath31 and @xmath29 at all values of @xmath33 .
there are additional regions which span out to large values of @xmath31 due to co - annihilations with light sleptons , particularly the lighter @xmath146 @xcite . at large @xmath33
, there are also regions in which the lightest neutralino sits on the s - channel pole of the pseudo - scalar and heavy scalar higgs producing ` funnel'-like regions @xcite
. finally , there are the so - called ` focus - point ' regions @xcite which are present at very large values of @xmath29 .
generally , these regions have a lower asymmetry ( because of the large value of @xmath29 ) , however , at values of @xmath147 , asymmetries as large as 10% are possible .
.2 in in fig .
[ fig1]a , we show the contours of constant @xmath148 in @xmath31@xmath3 plane for @xmath149 , @xmath150 and @xmath5 .
the constraints discussed above are shown by various curves and shaded regions .
the nearly vertical dashed ( black ) line at the left of the figure shows the chargino mass constraint .
allowed regions are to the right of this line .
the dot - dashed higgs mass contour ( red ) labeled 114 gev , always provides a stronger constraint .
allowed regions are again to the right of this curve .
however , one should be aware that there is a theoretical uncertainty in the higgs mass calculation , making this limit somewhat fuzzy .
the light ( violet ) solid curve shows the position of the 2-@xmath141 @xmath10 constraint , which again excludes small values of @xmath31 . in the dark ( red ) shaded region covering much of the upper left half of the plane ,
the lighter @xmath151 is either the lsp or is tachyonic .
since there are very strong constraints forbidding charged dark matter , this region is excluded .
the medium shaded ( green ) region shows the exclusion area provided by the @xmath152 measurements .
finally the light ( turquoise ) shaded region shows the area _ preferred _ by cosmology . outside this shaded region ,
the relic density is too small and is technically not excluded . putting all of the constraints together
, we find that for this value of @xmath153 gev and @xmath5 , the allowed region is bounded by @xmath154 and @xmath155 . in the allowed region
, the forward - backward asymmetry varies rapidly from very small ( unobservable ) values up to @xmath156 .
there is a wide region with an observable @xmath157@xmath156 asymmetry though the @xmath156 region is quite narrow ( restricted to @xmath158 ) . in fig .
[ fig1]b , we show the corresponding result for the opposite sign of @xmath4 .
while the cosmologically allowed region is qualitatively similar to the @xmath159 case and the higgs limit is slightly stronger , we see that the @xmath160 constraint is significantly stronger . indeed
, the combined constraints from @xmath160 and a @xmath146 lsp exclude @xmath161 for this value of @xmath29 and @xmath138 .
the 2-@xmath141 constraint from @xmath10 is also significantly stronger and when combined with the @xmath146 lsp constraint now exclude values of @xmath162 . as mentioned earlier , for @xmath138 both the sign and size of the asymmetry have changed . in general , the size of the asymmetry is suppressed by an order of magnitude .
clearly , the @xmath160 constraint now allows only a small region with a @xmath163 @xmath164 asymmetry . however , when all constraints are combined they exclude almost completely the otherwise allowed regions . at higher values of @xmath29 , slightly larger asymmetries are possible . at @xmath165 gev
( with @xmath138 ) , @xmath160 allows asymmetries as large as @xmath166% , however , the @xmath10 data still restricts the asymmetry to values below about @xmath167 % .
even at large @xmath33 and very large @xmath29 , we will see below that for @xmath138 , asymmetries never excced @xmath168 % . we note that independent of the sign of @xmath4 , the asymmetry is maximized for intermediate values of @xmath3 , @xmath169 it does not monotonically increase with increasing @xmath3 as was already argued earlier .
the main conclusion from this figure is that the sign of the @xmath4 parameter must be positive in order to have large observable @xmath170 .
we note that there are already @xmath8factory constraints due to recent belle and babar experiments @xcite .
for @xmath5 , they exclude a small region ( not plotted ) with @xmath171 and @xmath172 ( lying in the region with a charged lsp ) , whereas for @xmath6 the excluded region is shifted to @xmath173 and @xmath174 ( now lying in the region also excluded by @xmath175 ) .
the behaviour observed in fig . [
fig1]a ) and b ) is subject to large variations once the gut scale input parameters are varied .
this is seen in fig . [ fig1]c ) and d ) where the constant asymmetry curves are plotted in @xmath31@xmath3 for @xmath5 , @xmath149 , and @xmath176 and @xmath177 respectively .
it is clear that with increasing @xmath29 the cosmologically preferred ` bulk ' region is shifted towards larger @xmath3 values making it possible to get larger asymmetries .
in addition , we see very clearly the effect of @xmath178 coannihilations @xcite which extend the cosmological region to high values of @xmath31 .
the region below the ` bulk ' and coannihilation region is excluded as it corresponds to an area with @xmath179 .
while the chargino , higgs , and @xmath10 constraints are only slightly altered at the higher value of @xmath29 , we see that the charged lsp constraint is relaxed in fig .
[ fig1]c ) and greatly relaxed in panel d ) . for the higher values of @xmath29 we see from fig .
[ fig1]c ) that for @xmath180 and @xmath181 the forward - backward asymmetry ranges from @xmath182 to @xmath183 in regions which are not excluded by any experimental or cosmological constraints .
in particular , when @xmath184 and @xmath185 the asymmetry is well observable with a typical @xmath186 peak value .
here the @xmath8factory constraints are effective for @xmath187 and @xmath188 , i.e. only a small region in the upper left corner . for the @xmath177 case shown in fig .
[ fig1]d ) , the cosmologically allowed region is now shifted up past the maximum of @xmath189 , and is now typically @xmath156 . overall the forward - backward asymmetry is larger than 1% in the cosmologically allowed region and extends over the range @xmath190 and @xmath191 . we have also checked some cases with nonzero values of @xmath32 , assuming it to be either constant ( @xmath13 set to @xmath192 ) , or varying in proportion with @xmath31 ( @xmath13 , @xmath193 ) . for a variable @xmath32 , results were not qualitatively different from the results shown here . for a large and fixed value of @xmath32
, the comsological regions of interest could be very different @xcite , however , the asymmetry was found to be quantitatively similar as the results shown here .
however , we can not claim to have made a systematic examination of the @xmath194 parameter space .
.2 in in fig .
[ fig3 ] , we show the contours of @xmath195 ( left panel ) and @xmath196 ( right panel ) for @xmath149 , @xmath197 and @xmath5 . a comparison of the left panel with panel c ) of fig .
[ fig1 ] shows that there is very little difference between @xmath198 and @xmath199 final states as far as the asymmetry is concerned .
indeed , as mentioned before , the difference between the asymmetries is a measure of the su(3 ) flavor breaking or the different parameterizations of the associated form factors .
therefore , the similarity or dissimilarity of these two figures depends on how the hadronic effects are treated for the kaon and pion final states . on the other hand , the comparison between panel c ) of fig .
[ fig1 ] and the right panel of fig .
[ fig3 ] shows that the asymmetry is suppressed for the @xmath200 final states .
the asymmetry does not reach the @xmath182 level in any corner of the allowed regions .
.2 in in fig .
[ fig4]a ) , we show the @xmath170 contours as projected onto the @xmath31@xmath29 plane for @xmath149 , @xmath201 , @xmath5 .
the curves and shading in this figure are as in fig .
seen clearly are the ` bulk ' and coannihilation cosmological regions .
the latter traces the border of the @xmath146 lsp region which is now found at the lower right of the figure .
the constraints from @xmath160 and @xmath10 only exclude a small portion of the parameter plane at the lower left . in this figure , it is the area above the cosmological shaded region which is excluded due to an excessive relic density .
this figure illustrates the dependence of the asymmetries on the chargino and stop masses , and indeed , the asymmetry changes sign at fairly large @xmath29 and small @xmath31 .
this effect results in a sign change in @xmath23 due to the competition between the stop and chargino masses . as fig . [
fig4]a ) makes clear , the asymmetry is positive and remains below the @xmath202 level yielding essentially no observable signal at all .
an example with @xmath203 and @xmath138 is shown in fig . [ fig4]b ) .
while the cosmological region is similar to that in panel a ) , the constraint from @xmath160 is significantly stronger , as is the constraint from @xmath10 .
the latter imply that the asymmetry is never much larger than @xmath204 .
as one can see , it is difficult to get any observable effect at low values of @xmath3,for any sign of the @xmath4 parameter . in the lower two panels of fig .
[ fig4 ] , we show the contours of @xmath170 for larger values of @xmath33 . as one can see , the ` bulk ' cosmological regions are pushed to higher values of @xmath29 and we also see the appearance of the ` funnel ' regions where the lsp relic density is primarily controlled by @xmath205 s - channel annihilations . in panel c ) , there are large positive asymmetries ( @xmath206 ) in a broad region extending from @xmath207 all the way up to @xmath208 .
the @xmath8factory constraints only exclude a small region in the bottom left corner bounded by @xmath209 and @xmath210 .
panel d ) of fig .
[ fig4 ] , on the other hand , shows that , for negative @xmath4 and large @xmath33 , the allowed range of asymmetry is around @xmath164 never reaching the @xmath211 level mainly due to the @xmath212 constraint .
there are three cosmologically preferred strips , the wider one located in the region @xmath213 where the asymmetry is at most @xmath164 . for @xmath6
the @xmath8factory constraints exclude two small regions bounded by @xmath214 and @xmath215 , and @xmath216 and @xmath217 which both are already excluded by @xmath9 and cosmology .
we have discussed the forward backward asymmetry of @xmath218 decays which is generated by the flavor changing neutral current decays mediated by the higgs bosons .
in addition to the known properties @xmath13 the smallness of the muon production asymmetry compared to the @xmath2 production , the approximate independence of the asymmetry to the flavor of the final state meson , the reduction of the hadronic uncertainties in the high dilepton mass region , @xmath219 we find that * the remarkable enhancement of the asymmetry is a unique implication of susy not found in the sm and its two doublet version .
* the regions of large asymmetry ( @xmath206 ) always require @xmath5 , and when @xmath4 changes sign so does the asymmetry with an order of magnitude suppression in its size . *
the asymmetry in the decay channels with the lepton pair @xmath220 is significantly larger ( approximately by the factor @xmath221 ) than in the channels with @xmath200 .
thus in spite of a greater difficulty of restoring the kinematics in the events with @xmath2 leptons , these may still be advantageous for measuring the discussed asymmetry . *
the asymmetry is not a monotonically increasing function of @xmath3 instead it is maximized at intermediate values above which the scalar fcnc effects dominate and enhance the branching ratio , and below which such fcnc are too weak to induce an observable asymmetry . *
though @xmath175 strongly disfavors the negative values of @xmath4 , the present bounds from the muon @xmath10 measurement are much stronger , and it typically renders the asymmetry unobservably small . * the cosmological constraints are generally very restrictive . when @xmath222 gev , @xmath223 gev , @xmath224 and @xmath5 there is a relatively wide region where the asymmetry is @xmath186 for @xmath2 , and typically @xmath225 for muon final state .
* the @xmath8factory constraints are generally too weak to distort the regions of observable asymmetry , and the regions excluded by them are already disfavoured by one or more of @xmath175 , @xmath226 lsp and @xmath212 .
with increasing statistics it is expected that the branching ratios of semileptonic modes will be measured with better accuracy so that , for instance , the allowed range in ( [ belle ] ) will be narrowed . in this case , there may be useful constraints on regions of enhanced asymmetry .
this work is supported in part by the doe grant de - fg02 - 94er40823 .
99 m. s. alam _ et al . _ [ cleo collaboration ] , phys .
lett . * 74 * , 2885 ( 1995 ) ; s. chen _ et al . _
[ cleo collaboration ] , phys .
lett . * 87 * , 251807 ( 2001 ) [ arxiv : hep - ex/0108032 ] ; r. barate _
[ aleph collaboration ] , phys .
b * 429 * , 169 ( 1998 ) ; k. abe _ et al .
_ [ belle collaboration ] , phys .
b * 511 * , 151 ( 2001 ) [ arxiv : hep - ex/0103042 ] ; see also k. abe _ et al .
_ , [ belle collaboration ] , [ arxiv : hep - ex/0107065 ] ; l. lista [ babar collaboration ] , [ arxiv : hep - ex/0110010 ] .
b. aubert _ et al .
_ [ babar collaboration ] , phys . rev .
* 87 * , 091801 ( 2001 ) [ arxiv : hep - ex/0107013 ] ; k. abe _ et al . _
[ belle collaboration ] , phys .
lett . * 87 * , 091802 ( 2001 ) [ arxiv : hep - ex/0107061 ] . t. m. aliev and m. savci , phys .
b * 452 * , 318 ( 1999 ) [ arxiv : hep - ph/9902208 ] ; phys .
d * 60 * , 014005 ( 1999 ) [ arxiv : hep - ph/9812272 ] ; t. m. aliev , m. savci , a. ozpineci and h. koru , j. phys .
g * 24 * , 49 ( 1998 ) [ arxiv : hep - ph/9705222 ] .
lep higgs working group for higgs boson searches , opal collaboration , aleph collaboration , delphi collaboration and l3 collaboration , _
search for the standard model higgs boson at lep _ ,
aleph-2001 - 066 , delphi-2001 - 113 , cern - l3-note-2699 , opal - pn-479 , lhwg - note-2001 - 03 , cern - ep/2001 - 055 , arxiv : hep - ex/0107029 ; _ searches for the neutral higgs bosons of the mssm : preliminary combined results using lep data collected at energies up to 209 gev _ , lhwg - note-2001 - 04 , aleph-2001 - 057 , delphi-2001 - 114 , l3-note-2700 , opal - tn-699 , arxiv : hep - ex/0107030 . c. hamzaoui , m. pospelov and m. toharia , phys .
d * 59 * , 095005 ( 1999 ) [ arxiv : hep - ph/9807350 ] ; k. s. babu and c. kolda , phys . rev .
* 84 * , 228 ( 2000 ) [ arxiv : hep - ph/9909476 ] ; g. isidori and a. retico , jhep * 0111 * , 001 ( 2001 ) [ arxiv : hep - ph/0110121 ] .
g. degrassi , p. gambino and g. f. giudice , jhep * 0012 * , 009 ( 2000 ) [ arxiv : hep - ph/0009337 ] ; m. carena , d. garcia , u. nierste and c. e. wagner , phys .
b * 499 * , 141 ( 2001 ) [ arxiv : hep - ph/0010003 ] ; d. a. demir and k. a. olive , phys .
d * 65 * , 034007 ( 2002 ) [ arxiv : hep - ph/0107329 ] .
h. baer , m. brhlik , d. castano and x. tata , phys .
d * 58 * , 015007 ( 1998 ) [ arxiv : hep - ph/9712305 ] ; t. goto , y. okada , y. shimizu and m. tanaka , phys .
d * 55 * , 4273 ( 1997 ) [ arxiv : hep - ph/9609512 ] .
j. l. hewett and j. d. wells , phys .
d * 55 * , 5549 ( 1997 ) [ arxiv : hep - ph/9610323 ] .
p. h. chankowski and l. slawianowska , phys .
d * 63 * , 054012 ( 2001 ) [ arxiv : hep - ph/0008046 ] ; r. arnowitt , b. dutta , t. kamon and m. tanaka , arxiv : hep - ph/0203069 ; s. r. choudhury and n. gaur , phys .
b * 451 * , 86 ( 1999 ) [ arxiv : hep - ph/9810307 ] .
s. heinemeyer , w. hollik and g. weiglein , comput .
commun . * 124 * , 76 ( 2000 ) [ arxiv : hep - ph/9812320 ] ; s. heinemeyer , w. hollik and g. weiglein , eur .
j. c * 9 * ( 1999 ) 343 [ arxiv : hep - ph/9812472 ] .
m. knecht and a. nyffeler , arxiv : hep - ph/0111058 ; m. knecht , a. nyffeler , m. perrottet and e. de rafael , phys .
* 88 * , 071802 ( 2002 ) [ arxiv : hep - ph/0111059 ] ; m. hayakawa and t. kinoshita , arxiv : hep - ph/0112102 ; i. blokland , a. czarnecki and k. melnikov , phys .
lett . * 88 * , 071803 ( 2002 ) [ arxiv : hep - ph/0112117 ] ; j. bijnens , e. pallante and j. prades , nucl .
b * 626 * , 410 ( 2002 ) [ arxiv : hep - ph/0112255 ] .
l. l. everett , g. l. kane , s. rigolin and l. wang , phys .
lett . * 86 * , 3484 ( 2001 ) [ arxiv : hep - ph/0102145 ] ; j. l. feng and k. t. matchev , phys .
lett . * 86 * , 3480 ( 2001 ) [ arxiv : hep - ph/0102146 ] ; e. a. baltz and p. gondolo , phys .
lett . * 86 * , 5004 ( 2001 ) [ arxiv : hep - ph/0102147 ] ; u. chattopadhyay and p. nath , phys .
lett . *
86 * , 5854 ( 2001 ) [ arxiv : hep - ph/0102157 ] ; s. komine , t. moroi and m. yamaguchi , phys .
b * 506 * , 93 ( 2001 ) [ arxiv : hep - ph/0102204 ] ; j. ellis , d. v. nanopoulos and k. a. olive , phys .
b * 508 * ( 2001 ) 65 [ arxiv : hep - ph/0102331 ] ; r. arnowitt , b. dutta , b. hu and y. santoso , phys .
b * 505 * ( 2001 ) 177 [ arxiv : hep - ph/0102344 ] s. p. martin and j. d. wells , phys .
d * 64 * , 035003 ( 2001 ) [ arxiv : hep - ph/0103067 ] ; h. baer , c. balazs , j. ferrandis and x. tata , phys .
d * 64 * , 035004 ( 2001 ) [ arxiv : hep - ph/0103280 ] .
for other recent calculations , see , for example : j. r. ellis , t. falk , g. ganis and k. a. olive , phys . rev .
d * 62 * , 075010 ( 2000 ) [ arxiv : hep - ph/0004169 ] ; a. b. lahanas , d. v. nanopoulos and v. c. spanos , phys .
b * 518 * ( 2001 ) 94 [ arxiv : hep - ph/0107151 ] ; v. barger and c. kao , phys .
b * 518 * , 117 ( 2001 ) [ arxiv : hep - ph/0106189 ] ; l. roszkowski , r. ruiz de austri and t. nihei , jhep * 0108 * , 024 ( 2001 ) [ arxiv : hep - ph/0106334 ] ; a. djouadi , m. drees and j. l. kneur , jhep * 0108 * , 055 ( 2001 ) [ arxiv : hep - ph/0107316 ] .
h. baer , c. balazs and a. belyaev , arxiv : hep - ph/0202076 .
j. ellis , t. falk and k. a. olive , phys .
b * 444 * ( 1998 ) 367 [ arxiv : hep - ph/9810360 ] ; j. ellis , t. falk , k. a. olive and m. srednicki , astropart .
* 13 * ( 2000 ) 181 [ arxiv : hep - ph/9905481 ] ; m. e. gmez , g. lazarides and c. pallis , phys
. rev .
d * 61 * ( 2000 ) 123512 [ arxiv : hep - ph/9907261 ] and phys .
b * 487 * ( 2000 ) 313 [ arxiv : hep - ph/0004028 ] ; r. arnowitt , b. dutta and y. santoso , nucl .
b * 606 * ( 2001 ) 59 [ arxiv : hep - ph/0102181 ] .
m. drees and m. m. nojiri , phys . rev .
d * 47 * ( 1993 ) 376 [ arxiv : hep - ph/9207234 ] ; h. baer and m. brhlik , phys . rev .
d * 53 * ( 1996 ) 597 [ arxiv : hep - ph/9508321 ] and phys .
d * 57 * ( 1998 ) 567 [ arxiv : hep - ph/9706509 ] ; h. baer , m. brhlik , m. a. diaz , j. ferrandis , p. mercadante , p. quintana and x. tata , phys .
d * 63 * ( 2001 ) 015007 [ arxiv : hep - ph/0005027 ] ; a. b. lahanas , d. v. nanopoulos and v. c. spanos , mod .
a * 16 * ( 2001 ) 1229 [ arxiv : hep - ph/0009065 ]
. j. l. feng , k. t. matchev and t. moroi , phys .
* 84 * , 2322 ( 2000 ) [ arxiv : hep - ph/9908309 ] ; j. l. feng , k. t. matchev and t. moroi , phys .
d * 61 * , 075005 ( 2000 ) [ arxiv : hep - ph/9909334 ] ; j. l. feng , k. t. matchev and f. wilczek , phys .
b * 482 * , 388 ( 2000 ) [ arxiv : hep - ph/0004043 ] . | we analyze the forward backward asymmetry of the decays @xmath0 with @xmath1 or @xmath2 in the framework of the constrained minimal supersymmetric standard model .
we find that , the asymmetry is enhanced at large @xmath3 and depends strongly on the sign of the @xmath4 parameter . for @xmath5 ,
the asymmetry is typically large and observable whereas for @xmath6 , it changes the sign and is suppressed by an order of magnitude . including cosmological constraints we find that the asymmetry has a maximal value of about 30 % , produced when higgs- and gauge- induced flavor violations are of comparable size , at a value of @xmath7 .
the present constraints from the @xmath8factories are too weak to constrain parameter space , and the regions excluded by them are already disfavoured by at least one of @xmath9 , @xmath10 , and/or cosmology .
the size of the asymmetry is mainly determined by the flavor of the final state lepton rather than the flavor of the pseudoscalar . 6.5 in 8.7 in 0.0 in -0.5 in tpi
minn02/07 + umn th2047/02 + april 2002 * the forward
backward asymmetry of @xmath0 : supersymmetry at work * * d. a. demir , keith a. olive , * theoretical physics institute , university of minnesota , minneapolis , mn 55455 and + * m.b .
voloshin * theoretical physics institute , university of minnesota , minneapolis , mn 55455 + and + institute of theoretical and experimental physics , moscow 117218 |
several biological functions are triggered by mechanical force . these include stretching and contraction of muscle proteins such as titin @xcite , rolling and tethering of cell adhesion molecules @xcite , translocation of proteins across membranes @xcite , and unfoldase activity of chaperonins and proteasomes .
understanding these diverse functions requires probing the response of biomolecules to applied external tension .
dynamical responses to mechanical force can be used to characterize in detail the free energy landscape of biomolecules .
advances in manipulating micron - sized beads attached to single biomolecules have made it possible to stretch , twist , unfold and even unbind proteins using forces on the order of tens of piconewtons @xcite .
single molecule force spectroscopy on a number of different systems has allowed us to obtain a glimpse of the unbinding energy landscape of biomolecules and protein - protein complexes @xcite . in afm experiments ,
used to unfold proteins by force , one end of a protein is adsorbed on a template and a constant or a time - dependent pulling force is applied to the other terminus @xcite . by measuring the distribution of forces required to completely unfold proteins and the associated unfolding times , the global parameters of the protein energy landscape can be estimated @xcite .
these insightful experiments when combined with theoretical studies @xcite can give an unprecedented picture of forced - unfolding pathways .
current experiments have been designed primarily to obtain information on forced - unfolding of proteins and do not probe the reverse folding process .
although force - clamp afm techniques have been used recently to probe ( re)folding of single ubiquitin polyprotein @xcite , the lack of theoretical approaches has made it difficult to interpret these pioneering experiments @xcite .
secondly , the resolution of multiple timescales in protein folding and refolding requires not only novel experimental tools for single molecule experiments but also new theoretical analysis methods .
minimally , unfolding of proteins by a stretching force @xmath1 is described by the global unfolding time @xmath11 , timescales for propagation of the applied tension , and the dynamics describing the intermediates or `` protein coil '' states . finally , if the external conditions ( loading rate or the magnitude of @xmath1 ) are such that these processes can occur on similar timescales then the analysis of the data requires new theoretical ideas . for forced unfolding the variable conjugate to @xmath1 , namely , the protein end - to - end distance @xmath12 is a natural reaction coordinate .
however , @xmath12 is not appropriate for describing protein refolding which , due to substantial variations in the duration of folding barrier crossing , may range from milliseconds to few minutes . to obtain statistically _ meaningful _ distributions of unfolding times , a large number of _ complete _ unfolding trajectories must be recorded which requires repeated application of the pulling force .
the inherent heterogeneity in the duration of folding and the lack of correlation between evolution of @xmath12 and ( re)folding progress creates `` initial state ambiguity '' when force is repeatedly applied to the same molecule . as a result ,
the interpretation of unfolding time data is complicated especially when the conditions are such that reverse folding process at the quenched force @xmath4 can occur on a long timescale , @xmath13 .
motivated by the need to assess the effect of the multiple timescales on the energy landscape of folding and unfolding , we develop a new theoretical formalism to describe correlations between the various dynamical processes .
our theory leads naturally to a new class of single molecule force experiments , namely , the force correlation spectroscopy ( fcs ) which can be used to study both forced unfolding as well as force - quenched ( re)folding . such studies can lead to a more detailed information on both kinetic and dynamic events underlying unfolding and refolding . in the fcs ,
cycles of stretching ( @xmath1 ) are separated by periods @xmath3 of quenched force @xmath4@xmath5@xmath1 during which the stretched protein can relax from its unfolded state @xmath14 to coil state @xmath15 or even ( re)fold to the native basin of attraction ( nba ) state .
the two experimental observables are @xmath12 and the unfolding time @xmath16 .
the central quantity in the fcs is the distribution of unfolding times @xmath2 separated by recoil or refolding events of duration @xmath3 .
the higher order statistical measure embedded in @xmath2 is readily accessible by constructing a histogram of unfolding times for varying @xmath3 and does not require additional technical developments .
the crucial element in the proposed analysis is that @xmath2 is computed by _ averaging over final ( unfolded ) states _ , rather than initial ( folded ) states .
this procedure removes the potential ambiguity of not precisely knowing the initial distribution of conformations in the nba . despite the uniqueness of the native state
there are a number of conformations in the nba that reflect the fluctuations of the folded state .
the proposed formalism is a natural extension of unbinding time data analysis .
indeed , @xmath2 reduces to the standard distribution of unfolding times @xmath0 when @xmath3 exceeds protein ( re)folding timescale @xmath13 .
the complexity of the energy ladscape of proteins demands fcs and the theoretical analysis .
current single molecule experiments on poly - ub or poly - ig27 ( performed in the @xmath3@xmath17@xmath18 regime ) show that in these systems unfolding occurs abruptly in an apparent all - or - none manner or through a dominant intermediate @xcite . on the other hand ,
refolding upon force - quench is complex and surely occurs though an ensemble of collapsed coiled states @xcite .
a number of timescales characterize the stretch - release experiments . these include besides @xmath13 , the @xmath1-dependent unfolding time , and the relaxation dynamics in the coiled states @xmath19 upon force - quench @xmath20 .
in addition , if we assume that @xmath12 is an appropriate reaction coordinate then the location of the nba , @xmath19 , the transition state ensembles and the associated widths are required for a complete characterization of the underlying energy landscape .
most of these parameters can be extracted using the proposed fcs experiments and the theoretical analysis presented here . in a preliminary study @xcite
, we reported the basics of the theory which was used to propose a new class of single molecule force spectroscopy methods for deciphering protein - protein interactions .
the current paper is devoted to further developments in the theory with appplication to forced - unfolding and force - quench refolding of proteins .
in particular , we illustrate the efficacy of the fcs by analyzing single unfolding - refolding trajectories generated for a coarse - grained model ( cgm ) protein @xmath7 with @xmath8-sheet architecture @xcite .
we showed previously that forced - unraveling of @xmath7 , in the limit of @xmath3@xmath17@xmath18 , can be described by an apparent `` two - state ''
kinetics @xcite .
the thermodynamics and kinetics observed in @xmath7 is a characteristic of a number of proteins where folding / unfolding fits well two - state behavior @xcite .
thus , @xmath7 serves as a useful model to illustrate the efficacy of the fcs . here
, we show that by varying @xmath3 and the magnitude of the stretching ( @xmath1 or @xmath4 ) , the entire dynamical processes , starting from the nba to the fully stretched state , can be resolved . in the process
we establish that @xmath2 which can be measured using afm or laser optical tweezer ( lot ) experiments , provides a convenient way of characterizing the energy landscape of biomolecules in detail .
_ theory of force correlation spectroscopy ( fcs ) : _ in single molecule atomic force microscopy ( afm ) experiments used to unfold proteins by force , the n - terminus of a protein is anchored at the surface and the c - terminus is attached to the cantilever tip through a polymer linker ( figure 1 ) .
the molecule is stretched by displacing the cantilever tip and the resulting force is measured . from a theoretical perspective
it is more convenient to envision applying a constant stretching force @xmath1@xmath9@xmath21@xmath22 in the @xmath22-direction ( figure 1 ) .
the free energy in the constant force formulation is related to the experimental setup by a legendre transformation .
more recently , it has become possible to apply a constant force in afm or laser or optical tweezer ( lot ) experiments to the ends of a protein . with this setup the unfolding time for the end - to - end distance @xmath12 to reach the contour length @xmath23
can be measured for each molecule . for a fixed @xmath1 , repeated application of the pulling force results in a single trajectory of unfolding times ( @xmath24 , @xmath25 , @xmath26 , @xmath27 , figure 1 ) from which the histogram of unfolding times @xmath0 is obtained .
the @xmath1-dependent unfolding rate @xmath28 is obtained by fitting a poissonian formula @xmath29}$ ] to the kinetics of population of folded states @xmath30 which is related to @xmath0 as @xmath31@xmath9@xmath32@xmath33@xmath34 . because @xmath28 is a convolution of several microscopic processes , it does not describe unfolding in molecular detail . for instance , mechanical unfolding of fibronectin domains fniii involves the intermediate `` aligned '' state @xcite with partially disrupted hydrophobic core which can not be resolved by knowing only @xmath28 .
even when the transition from the folded state @xmath35 to the globally extended state @xmath36 @xcite does not involve parallel routes as in figure 2 , or multistate kinetics , the force - induced unfolding pathway must involve formation of intermediate coiled states @xmath19 . the subsequent transition from @xmath19 results in the formation of the globally unfolded state @xmath36 .
the incomplete time resolution prevents current experiments from probing the signature of the collapsed states . to probe the contributions from the underlying @xmath19 states to global unfolding requires sophisticated experiments that can resolve contributions from dynamic events underlying forced unfolding .
we propose a novel experimental procedure which , when supplemented with unfolding time data analysis described below , allows us to separately probe the kinetics of native interactions and the dynamics of the `` protein coil '' ( i.e. the dynamics of end - to - end distance @xmath12 when the native contacts are disrupted ) .
consider an experiment in which stretching cycles ( triggered by applying @xmath1 ) are interrupted by relaxation intervals @xmath3 during which force is quenched to @xmath4@xmath5@xmath1 . in the time interval @xmath3
, the polypeptide chain can relax into the manifold @xmath19 or even refold to the native state @xmath35 if @xmath3 is long enough . if @xmath1@xmath37@xmath38 and @xmath4@xmath5@xmath38 where @xmath38 is the equilibrium critical unfolding force at the specific temperature ( see phase diagram for @xmath7 in ref .
@xcite ) , these transformations can be controlled by @xmath3 . in the simplest implementation
we set @xmath4@xmath9@xmath10 .
the crucial element in the fcs experiment is that the same measurements are repeated for varying @xmath3 . in the fcs
the unfolding times are binned to obtain the joint histogram @xmath2 of unfolding events of duration @xmath16 generated from the recoil manifold @xmath39 or the native basin of attraction ( nba ) or both , depending on the duration of the relaxation time @xmath3 .
in the current experiments @xmath3@xmath17@xmath18 . as a result , the dynamics of additional states in the energy landscape that are explored during folding or unfolding are not probed .
the advantages of @xmath2 over the standard distribution of unfolding times @xmath0 are two - fold .
first , @xmath2 is computed by _ averaging over well - characterized fully stretched states_. this eliminates the problem of not knowing the distribution of initial protein states encountered in current experiments . indeed ,
due to intrinsic heterogeneity of the protein folding pathways , after the first unfolding event the protein may or may not refold into the native conformation , which creates the initial state ambiguity in the next ( second , third , etc ) pulling cycle .
therefore , statistical analysis based on averaging over final ( stretched ) states rather than initial ( folded ) states allows to overcome this difficulty .
secondly , statistical analysis of unfolding data _ performed for different values of @xmath3 _ allows us to separately probe the kinetics of native interactions and the dynamics of @xmath12 .
in addition , the entire energy landscape of native interactions can be mapped out when stretch - quench cycles are repeated for several values of @xmath1 , @xmath4 , and @xmath3 .
_ regime i _
( @xmath3@xmath40@xmath41 ) : in the simplest unfolding scenario application of @xmath1 results in the disruption of the native contacts ( @xmath35@xmath17@xmath19 ) followed by stretching of the manifold @xmath19 into @xmath36 ( figure 2 ) .
when stretching cycles are separated by short @xmath3 compared to the protein folding timescale @xmath41 at @xmath42@xmath9@xmath10 , @xmath43 is determined by the evolution of the coil state .
then the unfolded state population @xmath44 is given by the convolution of protein relaxation ( over time @xmath3 ) from the fully stretched state @xmath14@xmath45@xmath23 to an intermediate coiled state @xmath46 and streching @xmath46 into final state @xmath47 over time @xmath16 .
thus , @xmath43 is obtained from @xmath44 by taking the derivative with respect to @xmath16 , @xmath48 where @xmath49 is @xmath3-dependent normalization constant obtained by taking the last integral in the right hand side ( rhs ) of eq .
( [ 2.1 ] ) from @xmath47@xmath9@xmath10 to @xmath47@xmath9@xmath23 , and @xmath50 is the distribution of unfolded states .
if @xmath12 is well controlled , @xmath14 is expected to be centered around a fixed value @xmath51 and @xmath50@xmath52@xmath53 . in eq .
( [ 2.1 ] ) , @xmath54 and @xmath55 are respectively , the quenched and the stretching force dependent conditional probabilities to be in the coiled state @xmath56 at time @xmath16 arriving from state @xmath12 at time @xmath16@xmath9@xmath10 .
the integral over @xmath47 is performed in the range @xmath57 $ ] with @xmath12@xmath9@xmath23@xmath33@xmath58 ( figure 2 ) representing unfolding distance at which the total number of native contacts @xmath59 is at the unfolding threshold , @xmath59@xmath45@xmath60 .
it follows that @xmath43 ( eq .
( [ 2.1 ] ) ) contains information on the dynamics of @xmath12 . by assuming a model for @xmath12 and fitting @xmath43 , obtained by differentiating the integral expression appearing in eq .
( [ 2.1 ] ) , to the histogram of unfolding times , separated by short @xmath3@xmath40@xmath41 , we can resolve the dynamics of the polypeptide chain in the coil state which allows us to evaluate the @xmath42-dependent coil dynamical timescale @xmath61 using single molecule force spectroscopy .
the fit of eq .
( [ 2.1 ] ) could be analytical or numerical depending on the model of @xmath12 . _
regime ii _
( @xmath3@xmath62@xmath41 ) : when stretching cycles are interrupted by long relaxation periods , @xmath3@xmath62@xmath41 , the coiled states refold to @xmath63 ( figure 2 ) . in this regime ,
the initial conformations in forced - unfolding always reside in the nba . in this limit
, @xmath43 reduces to the standard distribution of unfolding times @xmath2@xmath17@xmath0 .
when @xmath3@xmath62@xmath41 , @xmath43 is given by the convolution of the kinetics of rupture of native contacts , resulting in protein extension @xmath64 , and dynamics of @xmath12 from state @xmath63@xmath65@xmath64 to final state @xmath47 , @xmath66 where @xmath67 is normalization constant obtained as in eq .
( [ 2.1 ] ) and @xmath68 is the probability of breaking the contacts over time @xmath16 that stabilize the native state @xmath63 . by assuming a model for @xmath68 and employing information on the dynamics of @xmath12 , obtained from the short @xmath3-experiment ( eq .
( [ 2.1 ] ) ) , we can probe the disruption kinetics of native interactions . by repeating long @xmath3-measurements at several values of @xmath1
, we can map out the energy landscape of native interactions projected on the direction of the end - to - end distance vector .
_ regime iii _ ( @xmath3@xmath52@xmath41 ) : in this limit , some of the molecules reach the nba , starting from extended states ( @xmath12@xmath45@xmath23 ) , whereas others remain in the basin @xmath19 .
the fraction of folding events @xmath69 depends on @xmath3 during which @xmath12 approaches the average extension @xmath70 facilitating the formation of native contacts .
thus , @xmath71 obtained in the intermediate @xmath3-experiment , involves contributions from both @xmath19 and @xmath35 initial conditions and is given by a superposition , @xmath72 where the probability to arrive to @xmath35 from @xmath19 at time @xmath3 is given by @xmath73 and the probability to remain in @xmath19 is @xmath74@xmath9@xmath32@xmath33@xmath75 . in eq .
( [ 2.5 ] ) , @xmath76 is the refolding probability determined by kinetics of formation of native contacts . because the dynamics of @xmath12 is weakly correlated with formation of native contacts , @xmath12 in @xmath77 is expected to be broadly distributed .
therefore , eqs .
( [ 2.4 ] ) and ( [ 2.5 ] ) can be used to probe kinetics of formation of native interactions . for eqs .
( [ 2.1 ] ) and ( [ 2.2 ] ) to be of use , one needs to know the ( re)folding timescale @xmath41 .
the simplest way to evaluate @xmath41 is to construct a series of histograms @xmath78 ( @xmath79@xmath9@xmath32 , @xmath80 , @xmath27 , @xmath81 ) for a fixed @xmath1 and increasing relaxation time @xmath82@xmath5@xmath83@xmath5@xmath27@xmath5@xmath84 , and compare @xmath78 s with the distribution @xmath85 obtained for sufficiently long @xmath86@xmath62@xmath41 . if @xmath3@xmath9@xmath86 then all the molecules are guaranteed to reach the nba .
the difference @xmath87 is expected to be non - zero for @xmath88@xmath89@xmath41 and should vanish if @xmath88 exceeds @xmath41 .
statistically , as @xmath88 starts to exceed @xmath41 increasingly more molecules will reach the nba by forming native contacts .
then , more unfolding trajectories will start from folded states , and when @xmath3@xmath62@xmath41 all unfolding events will originate from the nba .
therefore , @xmath90 is a sensitive measure for identifying the kinetic signatures for forming native contacts .
the utility of @xmath90 is that it is a simple yet accurate estimator of @xmath41 , which can be utilized in practical applications .
indeed , one can estimate @xmath41 by identifying it with the shortest @xmath88 at which @xmath91@xmath45@xmath85 , i.e. @xmath88@xmath45@xmath41 .
we should emphasize that to obtain @xmath41 from the criterion that @xmath92@xmath45@xmath10 no assumption about the distribution of refolding times have been made .
having evaluated @xmath41 one can then use eqs .
( [ 2.1 ] ) and ( [ 2.2 ] ) for short and long @xmath3-measurements to resolve protein coil dynamics and rupture kinetics of native contacts .
let us summarize the major steps in the fcs .
first , we estimate @xmath41 by using @xmath93 ( eq . [ 2.3 ] ) ) .
we next probe protein coil dynamics by analyzing @xmath94 obtained from short-@xmath3-measurements ( eq . (
[ 2.1 ] ) ) . in the third step , we use information on protein coil dynamics to resolve the kinetics of rupture of native interactions contained in @xmath95 of long-@xmath3-measurements ( eq .
( [ 2.2 ] ) ) . finally , by employing the information on protein coil dynamics and kinetics of rupture of native interactions , we resolve the kinetics of formation of native contacts by analyzing @xmath96 from intermediate @xmath3-measurements ( eqs .
( [ 2.4 ] ) and ( [ 2.5 ] ) ) .
the beauty of the proposed framework is that these experiments can be readily performed using available technology . in the current afm experiments
, @xmath3 can be made as short as few microseconds .
simple calculations show that the relaxation of a short @xmath97 amino acid protein from stretched state with @xmath23@xmath45@xmath98 to the coiled states @xmath19 with say , @xmath12@xmath45@xmath99 , occurs on the timescale @xmath61@xmath45@xmath100@xmath52@xmath101 , where @xmath102@xmath9@xmath23@xmath33@xmath12@xmath45@xmath103 and @xmath104@xmath45@xmath105 is the diffusion constant .
clearly , the time of formation of native contacts , which drives the transition from @xmath19 to the nba , prolongs @xmath41 by few microseconds to few miliseconds or larger , depending on folding conditions . in the experimental studies of forced unfolding and force - quenched refolding of ubiquitin
, @xmath41 was found to be of the order of @xmath106@xmath33@xmath107 @xcite .
computer simulation studies of unzipping - rezipping transitions in short @xmath108-nt rna hairpin p5ga have predicted that @xmath41 is of the order of few hundreds of microseconds @xcite . _
model for the kinetics of native contacts _ : to interpret the data generated by fcs it is useful to have a model for the time evolution of the native contacts and @xmath12 . we first present a simple kinetic model for rupture and formation of native contacts represented by probabilities @xmath109 and @xmath77 in eqs . ( [ 2.2 ] ) and ( [ 2.5 ] ) , respectively , and a model for the dynamics of @xmath12 given by the propagator @xmath110 . to describe the force - dependent evolution of native interactions
we adopt the continuous - time - random - walk ( ctrw ) formalism @xcite . in the ctrw model ,
a random walker , representing rupture ( formation ) of native contacts , pauses in the native ( coiled ) state for a time @xmath16 before making a transition to the coiled ( native ) state .
the waiting time distribution is given by the function @xmath111 ( @xmath112@xmath9@xmath113 or @xmath114 , where @xmath113 and @xmath114 refer to rupture and formation of native contacts , respectively ) .
we assume that the probabilities @xmath68 and @xmath115 are separable so that @xmath116 where @xmath117 is the equilibrium distribution of native states , @xmath118 is the distribution of coiled states and @xmath119 and @xmath120 are the force - dependent probabilities of rupture and formation of native contacts , respectively .
factorization in eq .
( [ 3.2 ] ) implies that application of force does not result in the redistribution of states @xmath63 and @xmath15 in the nba and in the manifold of coiled states @xmath19 , but only changes the timescales for nba@xmath17@xmath19 and @xmath19@xmath17nba transitions , and thus , the propabilities @xmath121 and @xmath122 .
we expect the approximation in eq .
( [ 3.2 ] ) to be valid provided the rupture of native contacts and refolding events are cooperative . during stretching cycles , for @xmath1 well above @xmath38
, we may neglect the reverse folding process .
similarly , global unfolding is negligible during relaxation periods with @xmath4@xmath5@xmath38 .
then , the master equations for @xmath123 is @xmath124 where @xmath125 is the generalized rate for the rupture and formation of native interactions . in the laplace domain , defined by @xmath126@xmath9@xmath127}$ ] , @xmath128 is related to @xmath125 as @xmath129^{-1}.\ ] ] the structure of the master equation for @xmath130 is identical to eq .
( [ 3.3 ] ) with the relationship between @xmath131 and @xmath132 being similar to eq .
( [ 3.4 ] ) .
the general solution to eqs .
( [ 3.3 ] ) is @xmath133^{-1}{\bar p}_{r}(0)\ ] ] where @xmath134@xmath9@xmath32 is the initial condition and the solution in the time domain is given by the inverse laplace transform @xmath135@xmath9@xmath136 .
the solution for @xmath137 is obtained in a similar fashion ( see eq .
( [ 3.5 ] ) ) with initial condition @xmath138@xmath9@xmath32 .
_ model for the polypeptide chain _ : in the extended state , when the majority of native interactions that stabilize the folded state are disrupted , the molecule can be treated roughly as a fluctuating coil .
simulations and analysis of native structures @xcite suggest that proteins behave as worm - like chains ( wlc ) .
for convenience we use a continuous wlc description for the coil state whose hamiltonian is @xmath139 + { \bf f}_{s , q}\int_{-l/2}^{l/2}ds \left({{\partial { \bf r}(s , t ) } \over { \partial s}}\right)\end{aligned}\ ] ] where @xmath6 is the protein coil persistence length .
a large number of force - extension curves obtained using mechanical unfolding experiments in proteins , dna , and rna have been analyzed using wlc model . in eq .
( [ 4.1 ] ) the three - dimensional cartesian vector @xmath140 represents the spatial location of the @xmath141 `` protein monomer '' at time @xmath16 .
the first two terms describe chain connectivity and bending energy , respectively .
the third term represents fluctuations of the chain free ends and the fourth term corresponds to coupling of @xmath142 to @xmath143 .
the end - to - end vector is computed as @xmath144@xmath9@xmath145@xmath33@xmath146 .
we need a dynamical model in which @xmath12 is represented by the propagator @xmath147 .
although bond vectors of a wlc chain are correlated , the statistics of @xmath12 can be represented by a large number of independent modes .
it is therefore reasonable , at least in the large @xmath23 limit , to describe @xmath148 by a gaussian , @xmath149}\ ] ] specified by the second moment @xmath150 and the normalized correlation function @xmath151@xmath9@xmath152 .
calculations of @xmath150 and @xmath151 are given in the appendix @xcite . in the absence of force ,
we obtain : @xmath153 where @xmath154 and @xmath155 are the eigenfunctions and eigenvalues of the modes of the operator that describe the dynamics of @xmath140 ( see eq .
( [ 4.3 ] ) ) . to construct the propagator @xmath148 for @xmath143 , eq .
( [ 4.3 ] ) is integrated with @xmath143 added to random force .
we obtain : @xmath156@xmath9@xmath157@xmath65@xmath158 , where @xmath159 .
we analyze the distributions of unfolding times @xmath2 for the model sequence @xmath7 ( figure 3 ) obtained using simulations , ctrw model for evolution of native interactions ( eqs .
( [ 3.2])-([3.5 ] ) ) and gaussian statistics of the protein coil ( eq . ( [ 4.2 ] ) ) .
_ simulations of model @xmath8-sheet protein : _ the usefulness of fcs is illustrated by computing and analyzing the distribution function @xmath43 for a model polypeptide chain with @xmath8-sheet architecture .
sequence @xmath7 , which is a variant of an off - lattice model introduced sometime ago @xcite , is a coarse - grained model ( cgm ) of a polypeptide chain , in which each amino acid is substituted with a united atom of appropriate mass and diameter at the position of the @xmath160-carbons @xcite .
the @xmath7 sequence is modeled as a chain of 46 connected beads of three types , hydrophobic @xmath161 , hydrophilic @xmath23 , and neutral @xmath81 , with the contour length @xmath162 , where @xmath163@xmath45@xmath164 is the distance between two consequtive @xmath165- carbon atoms .
the coordinate of @xmath166-th residue is given by the vector @xmath167 with @xmath166@xmath9@xmath32 , @xmath80 , @xmath27 , @xmath81 .
the potential energy @xmath36 of a chain conformation is @xmath168 where @xmath169 , @xmath170 , @xmath171 are the energy terms , which determine local protein structure , and @xmath172 corresponds to non - local ( non - bonded ) interactions .
the bond - length potential @xmath169 , that describes the chain connectivity , is given by a harmonic function @xmath173 where @xmath174@xmath9@xmath175@xmath176 and @xmath177 ( @xmath178 ) is the energy unit roughly equal to the free energy of a hydropobic contact .
the bending potential @xmath170 is @xmath179 where @xmath180@xmath9@xmath181@xmath182 and @xmath183@xmath9@xmath184 .
the dihedral angle potential @xmath171 , which is largely responsible for maintaining protein - like secondary structure , is taken to be @xmath185,\ ] ] where the coefficients @xmath186 and @xmath187 are sequence dependent .
along the @xmath8-strands @xmath188-states are preferred and @xmath189@xmath9@xmath161@xmath9@xmath190@xmath177 . in the turn regions ( i.e. in the vicinity of a cluster of @xmath81 residues )
@xmath189@xmath9@xmath10 , @xmath161@xmath9@xmath191@xmath177 .
the non - bonded 12 - 6 lennard - jones interaction @xmath172 between hydrophobic residues is the sum of pairwise energies @xmath192 where @xmath193 depend on the nature of the residues .
the double summation in eq .
( [ unb ] ) runs over all possible pairs excluding the nearest neighbor residues .
the potential @xmath194 between a pair of hydrophobic residues @xmath161 is given by @xmath195@xmath9@xmath196@xmath197@xmath198 @xmath199 $ ] , where @xmath197 is a random factor unique for each pair of @xmath161 residues @xcite and @xmath113@xmath9@xmath200@xmath33@xmath201 .
for all other pairs of residues @xmath202 is repulsive @xcite .
although an off - lattice cgm drastically simplifies the polypeptide chain structure , it does retain important charateristics of proteins , such as chain connectivity and the heterogeneity of contact interactions .
the local energy terms in s1 provide accurate representation of the protein topology .
the native structure of @xmath7 is a @xmath8-sheet protein that has a topology similar to the much studied immunoglobulin domains ( figure 3 ) .
when the model sequence is subject to @xmath1 or @xmath4 , the total energy is written as @xmath203@xmath9@xmath36@xmath33@xmath204@xmath205 ( @xmath112@xmath9@xmath206 or @xmath59 ) , where @xmath205 is the protein end - to - end vector , and @xmath143@xmath9@xmath207 is applied along the @xmath22-direction ( figure 1 ) .
the dynamics of the polypeptide chain is assumed to be given by the overdamped langevin equation , which in the absence of @xmath208 or @xmath42 , is @xmath209 where @xmath210 is the friction coefficient and @xmath211 is a gaussian white noise , with the statistics @xmath212 eqs .
( [ 2.14 ] ) are integrated with a step size @xmath213@xmath9@xmath214@xmath215 , where @xmath215@xmath9@xmath216@xmath9@xmath217 is the unit of time and @xmath218 is a residue mass . in eq .
( [ 2.14 ] ) the value of @xmath210@xmath9@xmath219 corresponds roughly to water viscosity .
_ simulations of unfolding and refolding of @xmath7 : _ for the model sequence @xmath7 we have previously shown that the equilibrium critical unfolding force is @xmath220@xmath221@xmath222 @xcite at the temperature @xmath223@xmath9@xmath224 below the folding transition temperature @xmath225@xmath9@xmath226 . at this temperature @xmath227@xmath228 of native contacts are formed ( see the phase diagram in ref .
@xcite ) . to simulate the stretch - relax trajectories ,
the initially folded structures in the nba were equilibrated for @xmath229 at @xmath223 .
to probe forced unfolding of s1 at @xmath3@xmath9@xmath223 , constant pulling force @xmath230@xmath9@xmath231@xmath232 and @xmath233 was applied to both terminals of @xmath7 . for these values of @xmath230 , @xmath7 globally unfolds in @xmath16@xmath9@xmath234 and @xmath235 , respectively .
cycles of stretching were interrupted by relaxation intervals during which the force is abruptly quenched to @xmath236@xmath9@xmath10 for various duration @xmath3 . unfolding - refolding trajectories of @xmath7 have been recorded as time series of @xmath12 and the number of native contacts @xmath59 . in figure 4
we present a single unfolding - refolding trajectory of @xmath12 and @xmath59 of @xmath7 , generated by stretch - relax cycles . stretching cycles of constant force @xmath230@xmath9@xmath233 applied for @xmath237
are interrupted by periods of quenched force relaxed over @xmath238 .
a folding event is registered if it results in the formation of @xmath239 of the total number of native contacts @xmath240@xmath9@xmath241 , i.e. @xmath59@xmath242@xmath243 for the first time .
an unfolding time is defined as the time of rupture of @xmath239 of all possible contacts for the first time . with this definition , the unfolded state end - to - end distance is @xmath12@xmath242@xmath14@xmath45@xmath244@xmath163 . in figure 4 ,
folded ( unfolded ) states correspond to minimal ( maximal ) @xmath12 and maximal ( minimal ) @xmath59 .
inspection of figure 4 shows that refolding events are essentially stochastic . out of 36 relaxation periods
only 9 attempts resulted in refolding of @xmath7 . both @xmath12 and @xmath59
show that refolding of @xmath7 occurs though an initial collapse to a coiled state with the end - to - end distance @xmath245@xmath45@xmath246 ( @xmath59@xmath45@xmath181 ) , followed by the establishment of additional native contacts ( @xmath59@xmath45@xmath247 ) stabilizing the folded state with @xmath248@xmath45@xmath249 .
we generated about 1200 single unfolding - refolding trajectories and monitored the time - dependent behavior of @xmath12 and @xmath59 .
in the first set of simulations we set @xmath21@xmath9@xmath250 and used several values of @xmath3@xmath9@xmath251 , @xmath252 , @xmath253 , @xmath254 and @xmath255 . in the second set , @xmath21@xmath9@xmath233 , and @xmath3@xmath9@xmath246 , @xmath256 , @xmath257 , @xmath258 and @xmath259 . each trajectory involves four stretching cycles separated by three relaxation intervals in which @xmath260@xmath9@xmath10 .
typical unfolding - refolding trajectories of @xmath12 and @xmath59 for @xmath21@xmath9@xmath250 , @xmath260@xmath9@xmath10 , and @xmath3@xmath9@xmath253 , @xmath254 and @xmath255 are displayed in figures 5 , 6 and 7 , respectively .
due to finite duration of stretching cycles ( @xmath238 ) , unfolding of @xmath7 failed in few cases which were not included in the subsequent analysis of unfolding times . only first stretching cycles in each trajectory are guaranteed to start from the nba and for @xmath3@xmath9@xmath261 ( figure 5 ) relatively few relaxation intervals result in refolding ( with large @xmath59 ) .
this implies that the distribution of unfolding times @xmath2 obtained from these trajectories are dominated by contributions from the coiled states with the kinetics of formation of the native contacts playing only a minor role .
not unexpectedly , refolding events are more frequent when @xmath3 is increased to @xmath262 and @xmath255 . at @xmath3@xmath9@xmath262
, @xmath59 reaches higher values ( @xmath45@xmath263 ) and the failure to refold is rare ( figure 6 ) .
this implies that as @xmath3 starts to exceed the ( re)folding time @xmath41 , the distribution of unfolding events , parametrized by @xmath2 , is characterized by diminishing contribution from the coiled states @xmath19 and is increasingly dominated by the folded conformations in the nba .
note that failed refolding events are observed even at @xmath3@xmath9@xmath255 ( figure 7 ) , which implies large heterogeneity in the duration of folding barrier crossing events . figures 5 - 7 suggest that the folding time @xmath41 at the temperature @xmath264 is in the range @xmath175@xmath33@xmath265 .
direct computations of the folding time @xmath41 from hundreds of folding trajectories starting with the fully stretched states gives @xmath266@xmath45@xmath267 .
the agreement between @xmath266 and @xmath41 validates our stretch - release simulations .
_ analysis of the distribution of unfolding times of @xmath7 : _ the theoretical considerations in our formalism suggest that the @xmath3-dependent heterogeneous unfolding processes occur not only from the nba but also from the intermediate coil @xmath19 states .
the @xmath3-dependent protein dynamics can be utilized to separately probe the coil dynamics of the polypeptide chain and the kinetics of formation / rupture of native contacts ( @xmath59 ) .
we now utilize unfolding - refolding trajectories of @xmath7 , simulated for short , intermediate and long @xmath3 , to build the histograms of unfolding times @xmath2 .
using @xmath2 we provide quantitative description of the polypeptide chain dynamics in the coil state and the kinetics of rupture and formation of native interactions by employing ctrw model for @xmath59 and gaussian statistics for @xmath12 .
we computed @xmath2 using the distribution of unfolding times obtained for @xmath21@xmath9@xmath233 , @xmath3@xmath9@xmath246 , @xmath256 and @xmath268 ( figure 8) , and @xmath21@xmath9@xmath250 , @xmath3@xmath9@xmath251 , @xmath252 and @xmath269 ( figure 9 ) . in both cases @xmath260@xmath9@xmath10 .
we excluded unfolding times corresponding to the first stretch - quench cycle of each trajectory which were used to construct @xmath0 for the purposes of comparing @xmath0 with @xmath2 for long @xmath3 .
single peaked @xmath2 obtained for @xmath3@xmath9@xmath270 ( figure 8) and @xmath3@xmath9@xmath271 ( figure 9 ) , represent contributions to @xmath7 unfolding from coil manifold @xmath39 alone .
when @xmath3 is increased to @xmath272 ( figure 8) and @xmath273 ( figure 9 ) , position of the peak shifts to longer times , i.e. from @xmath16@xmath45@xmath274 to @xmath16@xmath45@xmath275 ( figure 8) and from @xmath16@xmath45@xmath276 to @xmath16@xmath45@xmath277 ( figure 9 ) .
furthermore , @xmath2 develops a shoulder at @xmath16@xmath45@xmath277 and @xmath16@xmath45@xmath278 , observed for @xmath3@xmath9@xmath268 ( figure 8) and @xmath3@xmath9@xmath269 ( figure 9 ) , which indicates a growing ( with @xmath3 ) contribution to unfolding from relaxation trajectories that reach the nba . at longer @xmath3@xmath9@xmath262 , when most relaxation periods result in refolding of @xmath7 , contribution from coiled states diminishes and at @xmath3@xmath9@xmath265 @xmath2 is identical to the standard distribution @xmath0 constructed from unfolding times of the first stretch - quench cycle of each trajectory
this implies that for @xmath260@xmath9@xmath10 , @xmath41@xmath45@xmath265 and that @xmath2@xmath17@xmath0 for @xmath3@xmath37@xmath265 . the distribution @xmath2@xmath9@xmath0 constructed from unfolding times separated by @xmath3@xmath9@xmath279 is presented in figures 8 and 9
( top left panel ) .
we use the ctrw formalism to analyze the histograms of unfolding times @xmath2 from which the parameters that characterize the energy landscape of @xmath7 can be mapped .
we describe the kinetics of rupture and formation of native contacts by the waiting time distributions @xmath280 , @xmath281 , @xmath282 where @xmath283 ( dependent on @xmath21 ) and @xmath284 ( dependent on @xmath260 ) are the rates of rupture and formation of native interactions , respectively , @xmath285@xmath9@xmath286 are normalization constants ( @xmath287 is gamma function ) and @xmath288@xmath242@xmath32 are phenomenological parameters quantifying the deviations of the kinetics from a poissonian process .
for instance , @xmath288@xmath9@xmath32 implies poissonian process and corresponds to standard chemical kinetics with constant rate @xmath289 .
we assume that both the folded and the unfolded states are sharply distributed around the mean native and unfolded end - to - end distance @xmath290 and @xmath291 , respectively ( figure 2 ) , @xmath292 where @xmath293@xmath9@xmath244 residues corresponds to the definition of unfolded state . for @xmath7 the contour length @xmath294@xmath9@xmath295 .
thus , @xmath7 is unfolded if @xmath296 exceeds @xmath297 which implies @xmath298@xmath9@xmath106 residues ( see figure 2 and the lower limit of integration in eq .
( [ 2.1 ] ) ) .
we describe the distribution of states @xmath39 before the transition to the nba by a gaussian , @xmath299 with the width @xmath300 , centered around the average distance @xmath301 .
we performed numerical fits of the histograms presented in figures 8 and 9 using eqs .
( [ 2.1 ] ) , ( [ 2.2 ] ) , ( [ 2.4 ] ) and ( [ 2.5 ] ) . by fitting the theoretical curves to @xmath2 constructed from short @xmath3@xmath9@xmath270 and @xmath3@xmath9@xmath272 simulations ( figure 8) and @xmath3@xmath9@xmath271 and @xmath3@xmath9@xmath273 ( figure 9 ) , we first studied the dynamics of @xmath12 to estimate the dynamical timescale @xmath61 , i.e. the longest relaxation time corresponding to the smallest eigenvalue @xmath155 ( eq .
( [ 4.5 ] ) ) , and persistence length @xmath6 of @xmath7 in the coil states @xmath19 . by using the values of @xmath61 and @xmath6
, we used our theory to describe @xmath2 constructed from long @xmath3@xmath9@xmath279 simulations .
this analysis allows us to estimate the parameters characterizing the rupture of native contacts @xmath283 , @xmath302 , @xmath290 and @xmath64 .
finally , the parameters @xmath284 , @xmath303 , @xmath70 and @xmath300 , characterizing formation of native contacts were estimated using @xmath61 , @xmath6 , @xmath283 , @xmath302 , @xmath290 and @xmath64 , and fitting eqs .
( [ 2.4 ] ) and ( [ 2.5 ] ) to @xmath2 for intermediate @xmath3@xmath9@xmath268 ( figure 8) and @xmath3@xmath9@xmath269 ( figure 9 ) . _
extracting the energy landscape parameters of @xmath7 : _ there are a number of parameters that characterize the energy landscape and the dynamics of the major components in the nba@xmath17u transition .
the numerical values of the model parameters are summarized in the table .
the values of @xmath302@xmath9@xmath304 for @xmath21@xmath9@xmath250 and @xmath302@xmath9@xmath305 for @xmath21@xmath9@xmath233 indicate that rupture of native contacts is highly cooperative especially at the lower @xmath21@xmath9@xmath250 .
this agrees with the previous findings on kinetics of forced unfolding of @xmath7 @xcite which were based solely on unfolding @xmath7 by applying a constant force .
in contrast , the formation of native contacts is characterized by @xmath303@xmath45@xmath32 implying almost poissonian distribution for the kinetics of formation of native contacts .
the structural characteristics of the coil states are obtained using the relaxation of the polypeptide chain upon force - quench from stretched states .
the value of the persistence length @xmath6 , which should be independent of @xmath42 provided @xmath42@xmath306@xmath38@xmath40@xmath32 , is found to be about @xmath307@xmath308 ( table ) .
this value is in accord with the results of the recent experimental measurements based on kinetics of loop formation in denatured states of proteins @xcite . upon rupture of native contacts ,
the chain extends by @xmath309@xmath9@xmath310 ( for @xmath21@xmath9@xmath250 ) and @xmath309@xmath9@xmath311 ( for @xmath21@xmath9@xmath233 ) .
this distance separates the basins of folded states with @xmath312@xmath9@xmath313 at @xmath21@xmath9@xmath250 and @xmath312@xmath9@xmath314 at @xmath21@xmath9@xmath233 from high free energy states when the polypeptide chain is stretched in the direction of @xmath208 ( figure 2(a ) ) .
because these high free energy states are never populated we expect that forced - unfolding of @xmath7 must occur in an apparent two - step manner when @xmath3@xmath17@xmath18 .
explicit simulations of @xmath7 unfolding at constant @xmath208 ( @xmath45@xmath315 ) shows that mechanical unfolding occurs in a single step ( see figure 2 in ref .
@xcite ) . from the refolding free energy profile upon
force - quench ( see figure 2(b ) ) we infer that the initial stretched conformation must collapse to an ensemble of compact structures @xmath19 . from the analysis of @xmath43 using the ctrw formalism we find that the average end - to - end distance @xmath301 for the manifold @xmath19 is close to @xmath316 ( see the table ) which suggests that the ensemble of the @xmath19@xmath17 nba transition states is close to the native state .
there is a broad distribution of coiled states @xmath19 which is manifested in the large width @xmath317@xmath9@xmath318 . due to the broad conformational distribution
, there is substantial heterogeneity in the refolding pathways .
this feature is reflected in the long tails in @xmath2 ( see figures 8 and 9 ) . as a result , we expect the kinetic transition to be sharp .
the estimated timescale ( @xmath52@xmath319 ) for forming native contacts for @xmath7 is shorter than the coil dynamical timescale @xmath61 ( for the values of @xmath21 used in the simulations ) .
this indicates that the dynamical collapse of @xmath7 from the stretched state @xmath14@xmath45@xmath23 and equilibration in the coiled manifold @xmath19 constitites a significant fraction of the total folding time ( @xmath45@xmath61@xmath65@xmath320 ) . from the analysis of folding of @xmath7 ( @xmath43 at intermediate @xmath3 )
we also infer that the transition state ensemble for @xmath19@xmath17@xmath81 must be narrow . from the rates of rupture of native contacts @xmath283 at the two @xmath321 values and assuming bell model for the dependence of @xmath283 on @xmath21 , @xmath322 we estimated the force - free rupture rate @xmath323 and the critical extension @xmath324 at which folded states of @xmath7
become unstable .
we found that @xmath323@xmath9@xmath325@xmath326@xmath327 is negligible compared to the rate of formation of native contacts , @xmath284@xmath9@xmath328 .
the location of the transition state of unfolding @xmath12@xmath9@xmath290@xmath65@xmath324 is characterized by @xmath324@xmath9@xmath329@xmath163@xmath45@xmath330@xmath23 .
the value of @xmath324 is short compared to @xmath300 which is a measure of the width of the @xmath19 manifold .
small @xmath324 implies that the major barrier to unfolding is close to the native conformation .
a similar values of @xmath324 was obtained in the previous study of @xmath7 by using an entirely different approach @xcite .
these findings are consistent with afm experiments @xcite and computer simulations @xcite which show that native structures of proteins appear to be `` brittle '' upon application of mechanical force .
the parameter @xmath61 is an approximate estimate of the collapse time , @xmath331 , from the stretched to the coiled state . using direct simulations of the decay of the radius of gyration , @xmath332 ,
starting from a rod - like conformation , we obtained @xmath331@xmath45@xmath333 ( see supplementary information in @xcite ) .
the value of @xmath61 ( @xmath45@xmath334 ) is in reasonable agreement with the estimate of @xmath331 .
this exercise shows that reliable estimates of timescales of conformational dynamics , which are difficult to obtain , can be made using fcs . to ascertain the extent to which the estimate of @xmath28 agrees with independent calculations
, we obtained the @xmath28 by applying a constant force to unfold @xmath7 .
the value of @xmath28 , obtained by averaging over @xmath335 trajectories , is about @xmath336 at @xmath21@xmath9@xmath250 which is in rough accord with @xmath28@xmath45@xmath61@xmath65@xmath337@xmath45@xmath338 .
this further validates the efficacy of fcs in obtaining the energy landscape of proteins .
we also estimated @xmath339 from the value of @xmath28 obtained by direct simulation and the bell model .
the @xmath208-dependent unfolding rate @xmath28@xmath45@xmath61@xmath65@xmath337 increases with @xmath208 in accord with eq .
( [ 5.4 ] ) .
the prefactor ( @xmath339 ) is about ten fold smaller than @xmath323 .
the difference may be either due to the failure of the assumption that @xmath323@xmath9@xmath339 or the breakdown of the bell model @xcite .
in this section we summarize the main steps for practical implementation of the proposed force correlation spectroscopy ( fcs ) to probe the energy landscape of proteins using forced unfolding of proteins .
_ step 1 . evaluating the ( re)folding timescale @xmath41 : _ in the first phase of the fcs experiments
, one needs to collect a series of histograms @xmath78 , @xmath79@xmath9@xmath340 , @xmath81 of unfolding times for increasing relaxation time @xmath82@xmath5@xmath83@xmath5@xmath27@xmath5@xmath84 by repeated stretch - release experiments .
this can be done by discarding the first unfolding time @xmath24 in the sequence of recorded unfolding times @xmath341 for each @xmath88 to guarantee that all the unfolding events are generated from the stretched states with the distribution @xmath50 ( see eq .
( [ 2.1 ] ) ) .
this is a crucial element of the fcs methodology since it enables us to perform the averaging over the final ( stretched ) states .
it is easier to resolve experimentally the end - to - end distance @xmath12@xmath45@xmath23 , rather than the initial ( folded ) states in which a number of conformations belong to the nba .
the histograms are compared with @xmath85 obtained for sufficiently long @xmath86@xmath62@xmath41 . to ensure that @xmath86 exceeds @xmath41
, @xmath86 can be as long as few tens of minutes .
the time at which @xmath90 , given by eq .
( [ 2.3 ] ) , is equal to zero can then be used to estimate @xmath41 .
notice that our estimate of @xmath41 _ does not _ hinge on whether @xmath342@xmath343@xmath0 is poissonian or not !
clearly , the choice of @xmath86 depends on the protein under the study , and prior knowledge or bulk measurements of unfolding times observed under the influence of temperature jump or denaturing agents can serve as a guide to estimate the order of magnitude of @xmath86 .
_ step 2 . resolving the dynamics of the polypeptide chain : _ to this end
we have determined the ensemble average ( re)folding time , @xmath41 . in the second phase of the fcs
, we perform statistical analysis of the distribution of unfolding times collected at @xmath3@xmath40@xmath41 , i.e. @xmath94 ( regime i in section ii ) .
this allows us to probe the dynamic properties of the polypeptide chain , such as the protein persistence length @xmath6 and the protein dynamical timescale @xmath61 ( see the table ) .
indeed , by assuming a reasonable model for the conditional probability , @xmath344 , of the protein end - to - end distance and the distribution of the stretched states , @xmath50 , @xmath6 and @xmath61 can be determined from the fit ( either analytically or numerically ) of the unfolding time distribution , @xmath345 , given by eq .
( [ 2.1 ] ) , to the histogram of unfolding times collected for @xmath3@xmath40@xmath41 . to illustrate the utility of the fcs , in the present work we assumed a gaussian profile for @xmath110 ( see eq .
( [ 4.2 ] ) ) and the worm - like - chain model for the polypeptide chain . the general formulae ( [ 2.1 ] )
allows for the use of more sophisticated models of @xmath12 , should it become necessary .
recent single molecule fret experiments on proteins @xcite , dsdna , ssdna , and rna @xcite show , surprisingly , that the characteristics of unfolded states obey worm - like chain models .
moreover , all the data in forced unfolding of proteins have been analyzed using wlc models .
thus , the analysis of fcs data using wlc dynamics for unfolded polypeptide chains to a large extent is justified . @xmath55 and @xmath54 can be `` measured '' in the current afm and lot experiments by computing the frequency of occurence of the event @xmath12 after the forced stretch ( @xmath346@xmath9@xmath208 ) or force quench ( @xmath346@xmath9@xmath42 ) from the well - controlled partially stretched state @xmath12 or the fully stretched state @xmath12@xmath45@xmath23 of the chain , respectively , over time @xmath16 ( @xmath40@xmath41 ) .
_ step 3 . probing the kinetics of rupture of the protein native contacts : _ having resolved the dynamics of the protein in extension - time regime , where the number of native interactions that stabilize the native state is small
, we can resolve the kinetics of forced rupture of native interactions stabilizing the nba ( regime ii ) . in the third part of the fcs we analyze the distribution of unfolding times for @xmath3@xmath62@xmath41 , given by eq .
( [ 2.2 ] ) .
we use the knowledge about the propagator @xmath347 , appearing in the rhs of eq .
( [ 2.2 ] ) , obtained in _ step 2 _ to perform analytical or numerical fit of the distribution @xmath348 to the histogram of unfolding times collected for @xmath3@xmath62@xmath41 .
the new information , gathered in _ step 3 _ , sheds the light on the kinetics of native interactions stabilizing the nba , which is contained in the probability @xmath349 ( see eq .
( [ 2.2 ] ) ) . for convenience
, we used the continuous time random walk ( ctrw ) model for @xmath349 , which is summarized in eqs .
( [ 3.3])-([3.5 ] ) , and the assumption of separability , given by eqs .
( [ 3.2 ] ) .
ctrw reduces to the poissonian kinetics with the rate constants when the waiting time distribution function for the rupture of native contacts , @xmath128 , is an exponential function of @xmath16 .
the ctrw probes the possible deviations of the kinetics of @xmath349 from the poisson process and allows to test different functional forms for @xmath128 . in the simplest implementation of ctrw
utilized in the present work , @xmath128 is assumed to be an algebraic function of @xmath16 , given by eqs .
( [ 5.1 ] ) , which allows us to estimate the rate of rupture of native interactions , @xmath283 , and parameter @xmath302 quantifying the deviations of the rupture kinetics from a poissonian process .
furthermore , by repeating _
step 3 _ for different values of the stretching force , @xmath21 , and assuming the bell model for @xmath350 , given by eq .
( [ 5.4 ] ) , we can also estimate the force - free rupture rate , @xmath323 , and the critical extension @xmath324 , which quantifies the distance from the nba to the transition state along the direction of @xmath21 .
we also obtain the average end - to - end distance in the folded state , @xmath290 from the distribution of the native states @xmath117 .
_ step 4 . resolving the kinetics of formation of native interactions : _ in the final step
the distributions @xmath94 and @xmath348 , analyzed in _
steps 2 _ and _ 3 _ respectively , are used to form a linear superposition ( eq .
( [ 2.4 ] ) , regime iii ) .
the @xmath3-dependent weights are given by the probabilities @xmath74 and @xmath75@xmath9@xmath32@xmath33@xmath74 , respectively .
this superposition is used to fit the histogram of unfolding times , @xmath96 , collected for @xmath3@xmath52@xmath41 .
the estimated probability @xmath75 should then be matched with the propbability obtained by performing double integration in eq .
( [ 2.5 ] ) .
this allows us to probe the kinetics of formation of native contacts , @xmath351 , for the known propagator @xmath352 analyzed in _
step 2_. as in the case of @xmath349 , we assumed separability condition for @xmath353 ( eqs .
( [ 3.2 ] ) ) and ctrw for the kinetics of formation of native contacts contained in @xmath354 ( see eqs .
( [ 3.3])-([3.5 ] ) ) . a simple algebraic form for the waiting time distribution function , @xmath355 , given by eq .
( [ 5.1 ] ) , allows us to estimate the force - free rate of formation of native interactions , @xmath356@xmath9@xmath357 .
moreover , the heterogeneity of the protein folding pathyways can be assesses by analyzing the width , @xmath300 , of the distribution of coiled protein states , @xmath358 , centered around the average end - to - end distance , @xmath301 ( see eq .
( [ 5.3 ] ) ) .
similar to the analysis of rupture kinetics , _ step 4 _ could be repeated for the two values of the quenched force , @xmath260 , to yield the force - free rate of formation of native contacts , stabilizing the native fold , and the distance between @xmath301 and the transition state for the formation of native contacts . for the purposes of illustration , in the present work we used @xmath260@xmath9@xmath10 . at the minimum fcs can be used to obtain model - independent estimate of @xmath41 . by assuming a wlc description for coiled states , which is justified in light of a number of fret and forced unfolding experiments
, estimates of collapse times and their distribution as well as persistence length can be obtained .
if ctrw model is assumed then estimates of timescale for rupture and formation of native contacts can be made .
the utility of fcs for @xmath7 illustrates the efficacy of the theory .
the potential of obtaining hitherto unavailable information makes fcs extremely useful .
in this paper we have developed a theory to describe the role of internal relaxation of polypeptide chains in the dynamics of single molecule force - induced unfolding and force - quench refolding . to probe the effect of dynamics of the chain in the compact manifold of states , that are populated in the pathways to the nba starting from the stretched conformations , we propose using a series of stretch - release cycles . in this new class of single molecule experiments , referred to as force correlation spectroscopy ( fcs ) , the duration of release times ( @xmath3 ) is varied .
fcs is equivalent to conventional mechanical unfolding experiments in the limit @xmath3@xmath17@xmath18 . by applying our theory to a model @xmath8-sheet protein
we have shown that the parameters that characterize the energy landscape of proteins can be obtained using the joint distribution function of unfolding times @xmath43 .
the experimentally controllable parameters are @xmath1 , @xmath4 , and @xmath3 . in our illustrative example , we used values of @xmath1 that are approximately @xmath359 times greater than the equilibrium unfolding force .
we set @xmath4@xmath9@xmath10 which is difficult to realize in experiments . from the schematic energy landscape in figure 1
it is clear that the profiles corresponding to the positions of the manifold @xmath39 , the dynamics of @xmath39 , and the transition state location and barrier hight depend on @xmath4 .
the simple application , used here for proof of principle purposes only , already illustrates the power of fcs . to obtain the energy landscape of @xmath7 by using fcs that covers a broader range of @xmath1 and @xmath4 ,
a complete characterization of the landscape can be made .
the experiments that we propose based on the new theoretical development can be readily performed using presently available technology .
indeed , the pioneering experimental setup used by fernandez and li @xcite that have utilized force to initiate refolding can be readily adopted to perform single molecule fcs . it is known that even for proteins that fold in an apparent two - state manner the energy landscape is rough @xcite .
the scale of roughness @xmath360 can be measured in conventional afm experiments by varying temperature . the extent to which the internal dynamics of proteins is affected by @xmath360
, whose value is between @xmath361@xmath362 @xcite , on the force - quenched refolding is hard to predict .
these subtle effects of the energy landscape can be resolved ( in principle ) using fcs in which temperature is also varied .
in this appendix we outline the calculation of @xmath363 and @xmath364 for the force - free propagator @xmath365 . by using eq .
( [ 4.1 ] ) ( without the last term ) and applying the least action principle to wlc lagrangian @xmath23@xmath9@xmath366@xmath367@xmath368@xmath33@xmath369 , we obtain : @xmath370@xmath371@xmath140@xmath65@xmath372@xmath373 @xmath140@xmath33@xmath80@xmath374@xmath375 @xmath140@xmath9@xmath10 , where @xmath370 is the protein segment mass and @xmath372@xmath9@xmath376 , @xmath374@xmath9@xmath377 .
dynamics of the media is taken into account by including a stochastic force @xmath378 with the white noise statistics , @xmath379@xmath9@xmath10 , @xmath380@xmath9 @xmath80@xmath381@xmath382@xmath383@xmath384@xmath385 , where @xmath112@xmath9@xmath386 , @xmath387 , @xmath388 , and @xmath389 is the friction coefficient per unit coil length . in the overdamped limit ,
the equation of motion for @xmath140 is @xcite @xmath390 with the boundary conditions , @xmath391_{\pm l/2 } = 0 , \qquad \left [ 2\nu_0 { { \partial}\over { \partial s}}{\bf r}(s , t)+\epsilon { { \partial^2 } \over { \partial s^2}}{\bf r}(s , t ) \right]_{\pm l/2 } = 0\ ] ] where @xmath392@xmath9@xmath393 .
we solve eq .
( [ 4.3 ] ) by expanding @xmath140 and @xmath394 in a complete set of orthonormal eigenfunstions @xmath395 , i.e. @xmath396 substituting eqs . (
[ a.1 ] ) into eq .
( [ 4.3 ] ) and separating variables we obtain : @xmath397 where @xmath155 is the @xmath79-th eigenvalue . the second eq .
( [ a.2 ] ) for @xmath398 is solved by @xmath399}\ ] ] and the eigenfunctions @xmath400 are @xmath401 } } } \sin{[\alpha_n s ] } + { { \beta_n}\over { \cosh[\beta_n l/2 ] } } \sinh{[\beta_n s]}\right ) , n=1,3,\ldots , 2q+1 \\\nonumber \psi_n(s ) & = & \sqrt{{{c_n}/ { l}}}\left ( -{{\alpha_n}\over { \sin{[\alpha_n l/2 ] } } } \cos{[\alpha_n s ] } + { { \beta_n}\over { \sinh{[\beta_n l/2]}}}\cosh{[\beta_n s]}\right ) , n=2,4,\ldots , 2q\end{aligned}\ ] ] where @xmath402 s are the normalization constants ; @xmath403 and @xmath404 are determined from eqs .
( [ 4.4 ] ) , @xmath405}\cosh{[\beta_n l/2]}-\beta_n^3 \cos{[\alpha_n l/2 ] } \sinh{[\beta_n l/2 ] } \\\nonumber & - & { { 1}\over { l_p}}(\alpha_n^2+\beta_n^2)\cos{[\alpha_n l/2]}\cosh{[\beta_n l/2]}=0 , \quad n=1,3 , \ldots , 2q+1 \\\nonumber & \alpha_n & \cos{[\alpha_n l/2]}\sinh{[\beta_n l/2]}+\beta_n^3 \sin{[\alpha_n l/2 ] } \cosh{[\beta_n l/2 ] } \\\nonumber & + & { { 1}\over { l_p}}(\alpha_n^2+\beta_n^2)\sin{[\alpha_n l/2]}\sinh{[\beta_n l/2]}=0 , \quad n=2,4 , \ldots , 2q\end{aligned}\ ] ] the parameters @xmath403 and @xmath404 are related as @xmath406 .
the eigenvalues @xmath155 are given by @xmath155@xmath9@xmath372@xmath407@xmath65@xmath80@xmath374@xmath408 .
using eqs .
( [ a.1 ] ) and ( [ a.3 ] ) , we obtain : @xmath409 .
then , @xmath410@xmath9 @xmath411@xmath65 @xmath412@xmath33 @xmath413@xmath33 @xmath414 , which yields eq .
( [ 4.5 ] ) .
99 labeit , s. , and b. kolmerer .
titins : giant proteins in charge of muscle ultrastructure and elasticity .
_ science _ 270 : 293 - 296 .
minaeva , a. , m. kulke , j. m. fernandez and w. a. linke .
unfolding of titin domains explains the viscoelastic behavior od skeletal myofibrils .
_ biophys .
j. _ 80 : 1442 - 1451 .
ohashi , t. , d. p. kiehart and h. p. erickson .
dynamics and elasticity of the fibronectin matrix in living cell culture visualized by fibronectin - green fluorescent protein . _
usa _ 96 : 2153 - 2158 .
marshall , b. t. , m. long , j. w. piper , t. yago , r. p. mcever and c. zhu .
2003 . direct observation of catch bonds involving cell - adhesion molecules .
_ nature _ 423 : 190 - 193 .
evans , e. , a. leung , d. hammer and s. simon . 2001 .
chemically distinct transition states govern rapid dissociation of single l - selectin bonds under force . _
usa _ 98 : 3784 - 3789 .
barsegov , v. , and d. thirumalai . 2005 .
dynamics of unbinding of cell adhesion molecules : transition form catch to slip bonds .
usa _ 102 : 1835 - 1839 .
barsegov , v. , d. klimov and d. thirumalai .
langevin simulation studies of the transition from catch to slip bonds using coarse - grained protein models ( manuscript in preparation ) .
chicurel , m. e. , c. s. chen and d. e. ingber .
cellular control lies in the balance of forces .
cell biol .
_ 10 : 232 - 239 .
henrickson , s. e. , m. misakian , b. robertson and j. j. kasianowicz . 2000 .
driven dna transport into an asymmetric nanometer - scale pore .
_ 85 : 3057 - 3060 .
sung , w. , and p. j. park .
polymer translocation through a pore in a membrane .
_ 77 : 783 - 786 .
muthukumar , m. 2003 .
polymer escape through a nonapore .
_ j. chem .
_ 118 : 5174 - 5184 .
liphardt , g. , d. smith and c. bustamante . 2000 .
single - molecule studies of dna mechanics .
_ 10 : 279 - 285 .
allemand , j .- f . ,
d. bensimon and v. croquette .
2003 . stretching dna and rna to probe their interactions with proteins .
struct . biol .
_ 13 : 266 - 274 .
zhuang , x. , l. e. bartley , h. p. babcock , r. russel , t. ha , d. herschlag and s. chu . 2000 . a single molecule study of rna catalysis and folding .
_ science _ 288 : 2048 - 2051 .
rief , m. , h. clausen - shaunman and h. e. gaub .
sequence - dependent mechanics of single dna molecules .
_ 6 : 346 - 349 .
chang , k. c. , d. f. tees and d. a. hammer .
the state diagram for cell adhesion under flow : leukocyte rolling and firm adhesion .
usa _ 97 : 11262 - 11267 .
weisel , j. w. , h. shuman and r. i. litvinov .
protein - protein unbinding induced by force : single molecule studies .
_ 13 : 227 - 235 .
liphardt , j. , s. dumont , s. b. smith , i. jr tinoko and c. bustamante .
equilibrium information from nonequilibrium measurements in an experimental test of jarzynski s equality .
_ science _ 296 : 1832 - 1835 .
bartolo , d. , i. derenyi and a. ajdari . 2002 .
dynamic response of adhesion complexes : beyond the single path picture .
e _ 65 : 051910 - 051913 .
hummer , g. , and a. szabo .
free energy reconstruction from nonequilibrium single - molecule pulling experiments .
usa _ 98 : 3658 - 3661 .
lapidus , l. j. , w. a. eaton and j. hofrichter . 2000 . measuring the rate of intramolecular contact formation in polypeptides .
usa _ 97 : 7220 - 7225 .
hyeon , c. , and d. thirumalai .
2003 . can energy landscape roughness of proteins and rna be measured by using mechanical unfolding experiments ?
usa _ 100 : 10249 - 10253 .
oberhauser , a. f. , h. p. erickson and j. m. fernandez .
the molecular elasticity of the extracellular matrix protein tenascin .
_ nature _ 393 : 181 - 185
. fernandez , j. m. , and h. li .
force - clamp spectroscopy monitors the folding trajectory of a single protein . _ science _ 303 : 1674 - 1678 .
rief , m. , j. pascual , m. saraste and h. e. gaub .
single molecule force spectroscopy of spectrin repeats : low unfolding forces in helix bundles . _
_ 286 : 553 - 561 .
yang , g. , c. cecconi , w. a. baase , i. r. vetter , w. a. breyer , j. a. haack , b. w. matthews , f. w. dahlquist and c. bustamante . 2000 .
solid - state synthesis and mechanical unfolding of polymers of t4 lysozyme .
usa _ 97 : 139 - 144 .
craig , d. , a. krammer , k. schulten and v. vogel .
comparison of the early stages of forced unfolding for fibronectin type iii modules .
usa _ 98 : 5590 - 5595 .
carrion - vazquez , m. , a. f. oberhauser , s. b. fowler , p. e. marszalek , s. e. broedel , j. clarke and j. m. fernandez .
1999 . mechanical and chemical unfolding of a single protein : a comparison .
usa _ 96 : 3694 - 3600 .
scott , k. a. , a. steward , s. b. fowler and j. clarke .
titin : a multidomain protein that behaves as the sum of its parts .
_ 315 : 819 - 829 .
litvinovich , s. v. , and k. c. ingham .
1995 . interactions between type iii domains in the 110 kda cell - binding fragment of fibronectin .
_ 248 : 611 - 626 .
krammer , a. , h. lu , b. isralewitz , k. schulten and v. vogel .
1999 . forced unfolding of the fibronectin type iii module reveals a tensile molecular recognition switch .
natl . acad .
usa _ 96 : 1351 - 1356 .
isralewitz , b. , m. gao and k. schulten .
steered molecular dynamics and mechanical functions of proteins .
_ 11 : 224 - 230 .
cieplak , m. , t. x. hoang , m. o. robbins .
thermal effects in stretching of go - like models of titin and secondary structures .
_ proteins _ 56 : 285 - 297 .
makarov , d. e. , z. wang , j. b. thompson and h. g. hansma .
2002 . on the interpretation of force extension curves of single protein molecules .
_ j. chem .
_ 116 : 7760 - 7765 .
hyeon , c. , and d. thirumalai . 2005 .
chemical theory and computation special feature : mechanical unfolding of rna hairpins .
acad , sci .
usa _ 102 : 6789 - 6794 .
best , r. b. , and g. hummer . 2005 .
comment on `` force - clamp spectroscopy monitors the folding trajectory of a single protein '' .
_ science _ 308 : 498 .
barsegov , v. , and d. thirumalai .
2005 . probing protein - protein interactions by dynamic force correlation spectroscopy .
_ 95 : 168301 - 168305 .
hyeon , c. , r. i. dima and d. thirumalai .
size , shape , and flexibility of rna structures .
_ ( submitted ) .
honeycutt , j. d. , and d. thirumalai .
metastability of the folded states of globular proteins . _ proc .
acad , sci .
usa _ 87 : 3526 - 3529 .
klimov , d. k. , and d. thirumalai .
native topology determines force - induced unfolding pathways in globular proteins .
usa _ 97 : 7254 - 7259 .
veitshans , t. , d. k. klimov and d. thirumalai . 1997 .
protein folding kinetics : timescales , pathways and energy landscapes in terms of sequence - dependent properties . _ folding des . _ 2 : 1 - 22 .
jackson , s. e. 1998
. how do small single - domain proteins fold . _
folding and design _ 3 : r81-r91 .
montroll , e. w. , and h. sher .
anomalous transit - time dispersion in amorphous solids .
b _ 12 : 2455 - 2477 .
sher , h. , and m. lax .
stochastic transport in a disordered solid .
i. theory .
b _ 7 : 4491 - 4502 .
barsegov , v. , y. shapir and s. mukamel .
one - dimensional transport with dynamic disorder .
e _ 68 : 011101 - 011114 .
barsegov , v. , and s. mukamel .
probing single molecule kinetics by photon arrival trajectories .
_ j. chem .
_ 116 : 9802 - 9810 .
barsegov , v. , and s. mukamel .
multipoint fluorescence quenching - time statistics for single molecules with anomalous diffusion .
_ j. phys .
a _ 108 : 15 - 24 .
harnau , l. , r. g. winkler and p. reineker .
dynamic properties of molecular chains with variable stiffness . _ j. chem .
_ 102 : 7750 - 7757 .
dua , a. , and b. j. cherayil .
the dynamics of chain closure in semiflexible polymers . _ j. chem .
_ 116 : 399 - 409 .
lapidus , l. j. , p. s. steinbach , w. a. eaton , a. szabo and j. hofrichter .
effects of chain stiffness on the dynamics of loop formation in polypeptides .
appendix : testing a 1-dimensional diffusion model for peptide dynamics .
_ j. phys .
b _ 106 : 11628 - 11640 .
rief , m. , m. gautel , f. oesterhelt , j. m. fernandez and h. e. gaub .
reversible unfolding of individual titin immunoglobulin domains by afm .
_ science _ 276 : 1109 - 1112 .
klimov , d. k. , and d. thirumalai .
1999 . stretching single - domain proteins : phase diagram and kinetics of force - induced unfolding .
usa _ 96 : 6166 - 6170 .
li , m. s. , c. k. hu , d. klimov , d. thirumalai .
multiple stepwise refolding of immunoglobulin domain i27 upon force quench depends on initial conditions .
natl . acad .
( in press ) .
hummer , g. , and a. szabo .
kinetics from nonequilibrium single - molecule pulling experiments .
_ biophys .
j. _ 85 : 5 - 15 .
schuler , b. , e. a. lipman , p. j. steinbach , m. kumke and w. a. eaton .
polyproline and the `` spectroscopic ruler '' revisited with single - molecule fluorescence .
102 : 2754 - 2759 .
laurence , t. a. , x. kong , m. jaeger and s. weiss . 2005 . probing structural heterogeneities and fluctuations of nucleic acids and denatured proteins .
usa _ 102 : 17248 - 17353 .
caliskan , g. , c. hyeon , u. perez - salas , r. m. briber , s. a. woodson and d. thirumalai .
persistence length changes dramatically as rna folds .
( in press ) .
thirumalai , d. , and s. a. woodson .
kinetics of folding of proteins and rna .
_ acc . chem .
29 : 433 - 439 .
nevo , r. , v. brumfeld , r. kapon , p. hinterdorfer and z. reich .
direct measurement of protein energy landscape roughness .
_ embo reports _ 6 : 482 - 486 .
.energy landscape parameters for @xmath7 extracted from fcs : [ cols="^,^,^,^,^,^,^,^,^,^,^",options="header " , ] [ uctable.ps ]
* figure 1 . * @xmath163 : a typical afm setup : constant force @xmath346@xmath9@xmath1@xmath9@xmath21@xmath22 is applied through the cantilever tip linker in the direction @xmath22 parallel to the protein end - to - end vector @xmath205 .
stretching cycles are interrupted by relaxation intervals @xmath3 during which the force is quenched , @xmath346@xmath9@xmath4@xmath9@xmath260@xmath22 ( @xmath21@xmath37@xmath260 ) .
@xmath415 : a single trajectory of forced unfolding times @xmath24 , @xmath25 , @xmath26 , @xmath27 , separated by fixed relaxation time @xmath3 ,
during which the unfolded protein can either collapse into the manifold of colied states @xmath19 if @xmath3 is short or reach the native basin of attraction ( nba ) if @xmath3 is long . * figure 2 . *
schematic of the free energy profile of a protein ( red ) upon stretching at constant force @xmath208 and force - quench @xmath42 .
( a ) : the projections of energy landscape ( blue ) is in the direction of @xmath205 which is a suitable reaction coordinate for unfolding induced by force @xmath208 .
the average end - to - end distance in the native basin of attraction is @xmath316 . upon application of @xmath1 , rupture of contacts that stabilize the folded state @xmath35 results in the formation of an ensemble of high energy extended ( by @xmath64 ) conformations @xmath416 .
subsequently , transitions to globally unfolded state @xmath36 ( with @xmath417@xmath89@xmath12@xmath89@xmath23 ) occurs .
( b ) : free energy profile for force - quench refolding which occurs in the order @xmath36@xmath17@xmath19@xmath17@xmath35 .
refolding is initiated by quenching the force @xmath208@xmath17@xmath42@xmath5@xmath38 , where @xmath38 is the equilibrium critical force needed to unfold the native protein .
the initial event in the process is the formation of an ensemble of compact structures .
the mean end - to - end distance of @xmath19 is @xmath301 and the width is @xmath300 which is a measure of heterogeneity of the refolding pathways .
these states may or may not end up in the native basin of attraction ( nba ) depending on the duration of @xmath3 .
we have used @xmath205 as a reaction coordinate during force - quench for purposes of illustration only .
native structure of the model protein @xmath7 .
the model polypeptide chain has a @xmath8-sheet architecture of the native state .
the @xmath8-strands of the model chain are formed by native contacts between hydrophobic residues ( given by blue balls ) .
the hydrophilic residues are shown by red balls and the residues forming the turn regions are given in grey .
* figure 4 . * a single unfolding - refolding trajectory of the end - to - end distance @xmath296 ( black ) and the total number of native contacts @xmath59 ( red ) as a function of time @xmath16 for @xmath7 .
the trajectory is obtained by repeated application of stretch - quench cycles with stretching force @xmath230@xmath9@xmath418 and quenched force @xmath260@xmath9@xmath10 .
the duration of streching cycle and relaxation period is @xmath419 and @xmath336 , respectively .
the first five unfolding events corresponding to large @xmath296 and small @xmath59 are marked explicitely by numbers @xmath32 , @xmath80 , @xmath420 , @xmath196 and @xmath421 .
force stretch and force quench for the stretch - quench cycles @xmath422 , @xmath423 , @xmath246 , @xmath424 and @xmath425 ( middle panel ) are denoted by solid green and dash.dotted blue arrows . * figure 5 . * typical unfolding - refolding trajectories of @xmath296 ( black ) and @xmath59 ( red ) for @xmath7 as functions of time @xmath16 , simulated by applying four stretch - quench cycles at the pulling force @xmath230@xmath9@xmath426 and quenched force @xmath260@xmath9@xmath10 . the duration of relaxation time @xmath3@xmath9@xmath269 .
* figure 6 .
* examples of unfolding - refolding trajectories of @xmath296 ( black ) and @xmath59 ( red ) for @xmath7 as a function of time @xmath16 .
the pulling force is @xmath230@xmath9@xmath426 and the quenched force is @xmath260@xmath9@xmath10 .
the duration of relaxation time @xmath3@xmath9@xmath262 .
* figure 8 .
* histograms of forced unfolding times @xmath0 and the joint distributions of unfolding times separated by relaxation periods of the quenched force @xmath2 .
the distribution functions are constructed from single unfolding - refolding trajectories of @xmath7 simulated in stretch - quench cycles of @xmath21@xmath9@xmath233 and @xmath260@xmath9@xmath10 for @xmath3@xmath9@xmath427 , @xmath428 and @xmath429 .
simulated distributions are shown by red bars with the contribution to global unfolding events from coiled conformations @xmath19 indicated by an arrow for @xmath3@xmath9@xmath429 .
the results of the numerical fits obtained by using eqs .
( [ 2.1])-([2.5 ] ) are represented by solid lines .
the energy landscape parameters of @xmath7 are summarized in the table .
* figure 9 .
* histograms of forced unfolding times @xmath0 and @xmath2 constructed from single unfolding - refolding trajectories for @xmath7 .
the stretch - quench cycles were simulated with @xmath21@xmath9@xmath250 and @xmath260@xmath9@xmath10 for @xmath3@xmath9@xmath430 , @xmath431 and @xmath261 .
simulated distributions are shown by red bars with the contribution to global unfolding events from coiled conformations @xmath19 indicated by an arrow for @xmath3@xmath9@xmath261 .
the results of numerical fit obtained by using eqs .
( [ 2.1])-([2.5 ] ) are represented by solid lines .
the values of the parameters are given in the table . | we present a new theory that takes internal dynamics of proteins into account to describe forced - unfolding and force - quench refolding in single molecule experiments . in the current experimental setup ( atomic force microscopy or laser optical tweezers )
the distribution of unfolding times , @xmath0 , is measured by applying a constant stretching force @xmath1 from which the apparent @xmath1-dependent unfolding rate is obtained . to describe the complexity of the underlying energy landscape requires additional probes that can incorporate the dynamics of tension propagation and relaxation of the polypeptide chain upon force quench .
we introduce a theory of force correlation spectroscopy ( fcs ) to map the parameters of the energy landscape of proteins . in the fcs
the joint distribution @xmath2 of folding and unfolding times is constructed by repeated application of cycles of stretching at constant @xmath1 separated by release periods @xmath3 during which the force is quenched to @xmath4@xmath5@xmath1 . during the release period
, the protein can collapse to a manifold of compact states or refold .
we show that @xmath2 at various @xmath1 and @xmath4 values can be used to resolve the kinetics of unfolding as well as formation of native contacts .
we also present methods to extract the parameters of the energy landscape using chain extension as the reaction coordinate and @xmath2 .
the theory and a worm - like chain model for the unfolded states allows us to obtain the persistence length @xmath6 and the @xmath4-dependent relaxation time , that gives an estimate of collapse timescale at the single molecular level , in the coil states of the polypeptide chain .
thus , a more complete description of landscape of protein native interactions can be maped out if unfolding time data are collected at several values of @xmath1 and @xmath4 .
we illustrate the utility of the proposed formalism by analyzing simulations of unfolding - refolding trajectories of a coarse - grained protein ( @xmath7 ) with @xmath8-sheet architecture for several values of @xmath1 , @xmath3 and @xmath4@xmath9@xmath10 .
the simulations of stretch - relax trajectories are used to map many of the parameters that characterize the energy landscape of @xmath7 . |
direct numerical simulations ( dns ) of the equations governing turbulent phenomena have now become an important tool in both research and applications .
faithful results require the equations to be solved on a grid of adequate resolution .
most solutions of the past adopt a standard resolution based on an average measure of the smallest scale estimated from dimensional considerations .
our purpose here is to show , for the case of turbulent mixing , that such standard measures do not resolve a range of very fine scales that are present in reality .
we demonstrate the essentials of this feature by performing several numerical simulations whose resolution is much finer than that adopted routinely , and comment on the practical relevance of the scales missed in past calculations .
we focus on the practically important case of @xmath0 , where @xmath1 is the schmidt number , @xmath2 the kinematic viscosity of the fluid and @xmath3 the scalar diffusivity .
the smallest scale of mixing is generated by the balance of turbulent stretching and diffusion ( batchelor 1959 ) , and is related to the kolmogorov scale @xmath4 through the identity @xmath5 ; here , @xmath6 is the average value of the instantaneous ( and local ) energy dissipation rate @xmath7 of the turbulent kinetic energy .
the expression for @xmath7 reads as @xmath8 where @xmath9 is the velocity fluctuation in the direction @xmath10 . in simulations of a fixed size , in which the advection diffusion equation and the navier - stokes equations are simultaneously solved ,
the largest scale has traditionally been maximized by resolving no more than @xmath11 .
resolving this scale is also the goal of most experimental efforts , though they often fall short for large pclet numbers @xmath12 , where @xmath13 and @xmath14 , respectively , are the characteristic large - scale velocity and length scale of the flow .
however , since the time it was realized that the local energy dissipation rate @xmath7 has a multifractal character ( e.g. , sreenivasan & meneveau 1988 ) , and that its average value @xmath6 is therefore not adequate for quantifying small - scale characteristics ( grant , stewart & moilliet 1962 ) , it has been clear that , locally , length scales smaller than @xmath15 appear in a flow .
this is obvious from the definition of the _ local _ kolmogorov scale @xmath16 , in which @xmath17 is a highly fluctuating variable : the larger the value of @xmath7 locally , the smaller is the local kolmogorov scale @xmath18 , and vice versa . in view of the relation between the _ local _ values of the batchelor and kolmogorov scales , namely @xmath19 ,
it becomes similarly apparent that the existence of very small values of @xmath18 may lead to @xmath20 values that are much smaller than the average value @xmath11 .
this intuitive ( and incomplete ) reasoning will be supplemented by further explanation momentarily , but it is enough to note here that such fine scales have not been explored before .
section 2 presents a brief description of the simulations and resolution effects , and the main results are stated in section 3 .
this is followed by a discussion in section 4 of how the finest dissipation filaments are generated .
conclusions are stated in section 5 . and
@xmath21 , as indicated in the legend . in ( b ) we have fitted the data by @xmath22 where @xmath23 , with @xmath24 and @xmath25 as in chen _
( 1993 ) . ]
the advection - diffusion equation for the passive scalar is solved together with the navier - stokes equations for the homogeneous and isotropic turbulent velocity field .
the simulations are carried out within periodic boxes of fixed size @xmath26 .
the flow is maintained stationary by a random volume forcing .
the passive scalar is rendered stationary through a mean scalar gradient in the @xmath27-direction . except for enhanced resolution , about which some comments follow , the pseudo - spectral methods used here are quite standard ( schumacher & sreenivasan 2003 ; vedula , yeung & fox 2001 )
. for dns of box - type turbulence on an @xmath28 grid based on pseudo - spectral methods control of alias errors requires that modes with wavenumber higher than @xmath29 be truncated ( patterson & orszag 1971 ) .
specifically , this removes double and triple aliases in three dimensions .
the usual resolution criterion is expressed as a number for @xmath30 which , by extension to cases with @xmath31 , is @xmath32 a proper resolution of the smallest scale requires that the global minimum of @xmath20 be resolved by the grid which can be written as @xmath33}{\delta}\ge 1\ , .
\label{criterion1}\ ] ] by comparing ( 2.2 ) with the usual resolution criterion ( [ criterion0 ] ) , we get @xmath34 where we used @xmath35 and @xmath29 .
it is apparent that the criterion ( [ criterion1 ] ) is stronger
. our simulation will satisfy ( [ criterion1 ] ) and its results will be compared with data from those with nominal grid resolution .
unfortunately , the demands of this fine resolution restrict us to rather low values of the reynolds number ( @xmath36 , see table 1 for details ) .
this necessitates that most of our flow scales are in the viscous range of turbulence .
figure [ spec](a ) shows the energy spectrum for @xmath37 , evaluated with two resolutions , @xmath38 and 512 .
it is seen that the noise floor is reached some 30 orders of magnitude below the peak signal . in figure [ spec](b ) , we verify that the energy spectrum in the far - dissipation range falls off as @xmath22 with @xmath23 . for the lower reynolds number case , one gets @xmath24 and @xmath25 as did chen _ et al . _
( 1993 ) for the same range of dissipative wavenumbers ; this range is indicated by the two vertical dotted lines in the figure .
.parameters of the numerical simulations .
@xmath39 is the ( equidistant ) grid spacing .
the integral length scale for the velocity field is @xmath40 and that for the scalar field is @xmath41 ; here @xmath42 and @xmath43 are the spectral densities of the velocity and scalar fields , respectively , @xmath44 is the mean square of the velocity in the @xmath45-direction , and @xmath46 that of the passive scalar .
@xmath47 is the averaging time in units of the large scale eddy turnover time @xmath48 . [
cols="<,^,^,^",options="header " , ] [ tab2 ] of particular interest is the nature of the velocity field near the most intense or the least intense parts of the scalar dissipation .
this was examined through the eigenvalue analysis of the velocity gradient at sites where @xmath49 exceeded or fell below a chosen threshold .
trends with @xmath1 were discussed by schumacher & sreenivasan ( 2003 ) , but here we focus only on the highest schmidt number . the results of table 2 show that the pure straining motion ( three real eigenvalues ) and rotation ( complex conjugated pair of eigenvalues ) contribute to low amplitude scalar dissipation . it may be thought that the scale @xmath50 of maximum dissipation is related closely to the most compressive strain events , i.e. @xmath51 .
the present data for all levels of @xmath52 , some of which are reported in table 2 , indicate that this is not so .
indeed , a broad range of compressive rates not merely the maximal magnitudes are associated with intense scalar dissipation .
this is already apparent from the analysis of ashurst _
( 1987 ) ( see their figure 9b which reports an isotropic flow for sc=0.5 ) . while the most compressive eigenvector was preferentially aligned with the direction scalar gradient , corresponding events of maximum @xmath53 were preferentially aligned at an angle of about @xmath54 .
we explicitly note that this is the eulerian point of view , and that a lagrangian analysis following incipient fronts to the stage of their maturity may yield a different result .
the possibility that the resolution requirement could be more stringent than is conventionally believed has been discussed to some detail in sreenivasan ( 2004 ) , but the details outlined in the present paper have not been explored before .
when there is a significant overlap of the intermediate dissipation with the viscous - convective range , extreme values of the scalar dissipation are determined by the roughest velocity increments of the inertial range of turbulence .
an important example in which these fine scales would make a difference is non - premixed turbulent combustion ( bilger 2004 ) . there , the scalar dissipation rate of the mixture fraction enters as a basic quantity , e.g. for the modelling of jet - diffusion flames .
chemical reactions take place at the stochiometric mixture fraction in sheets of sub - kolmogorov thickness .
one expects steep gradients across such layers and strongly varying scalar dissipation .
these variations are not captured in the flamelet equations where only the statistical mean enters the expansion parameter ( peters 2000 ) .
a broad example where resolution effects can be important is the multifractal scaling of @xmath7 and @xmath53 .
the spiky structures in space and time , which are most prominent in the dissipative scales , are thought to affect appropriate turbulent field even for scales larger than @xmath18 or @xmath20 , as appropriate ( frisch 1995 ) . to summarize
, the very fine scalar filaments that were resolved here do not seem to be associated with the most intense scalar dissipation .
the small - scale stirring in the flow seems to interrupt a further steepening process which is known as the formation of mature fronts .
similar findings were made for two - dimensional turblence at @xmath55 ( celani _ et al .
clearly , the reynolds number of the advecting flow has an impact on this issue simply because the intermediate dissipation range might `` overshadow '' the viscous - convective range completely for sufficiently large @xmath56 .
a more conclusive study of this issue will require higher reynolds numbers and will be part of future work .
+ computations were carried out using the npaci resources provided by the san diego supercomputer center and on the jlich multiprocessor ( jump ) ibm cluster at the john von neumann - institute for computing , jlich .
special thanks go to herwig zilken ( jlich ) for his help with figure [ volume ] .
pky and krs acknowledge support by the us national science foundation and js from the deutsche forschungsgemeinschaft . | we explore very fine scales of scalar dissipation in turbulent mixing , below kolmogorov and around batchelor scales , by performing direct numerical simulations at much finer grid resolution than is usually adopted in the past .
we consider the resolution in terms of a local , fluctuating batchelor scale and study the effects on the tails of the probability density function and multifractal properties of the scalar dissipation .
the origin and importance of these very fine - scale fluctuations are discussed .
one conclusion is that they are unlikely to be related to the most intense dissipation events . |
in the forthcoming neutrino experiments , @xmath1 appearance which is the positive evidence of @xmath2 oscillation is the one of main subjects , as well as precise measurements of neutrino oscillation parameters and unknown @xmath3 or cp phase measurement . in 2005
, icarus @xcite and opera @xcite experiments using the cngs beam @xcite will start to detect @xmath1 appearance .
@xmath4 should be detected through the @xmath5 production by charged current reactions off a nucleon target .
as we pointed out in the recent paper @xcite , the information on the spin polarization of @xmath5 produced by neutrino is essential to determine the @xmath5 production signal since the decay particle distributions depend crucially on the @xmath5 polarization .
we present polarization effects in @xmath5 production by neutrino - nucleon scattering via charged current .
we consider quasi - elastic scattering ( qe ) , @xmath0 resonance production ( res ) and deep inelastic scattering ( dis ) processes , which give dominant contributions in the cngs beam energy region , in the laboratory frame @xcite .
we show the physical regions of kinematical variables and give the relation between the @xmath5 spin polarization vector and the spin density matrix of the charged current @xmath5 production process . in this report , we take account of only @xmath6 production
. you can find more details in ref .
@xcite .
we consider @xmath6 production for each process by neutrino off a nucleon target in the laboratory frame ; @xmath7 the four - momenta are @xmath8 and lorentz invariant variables are defined as @xmath9 each process is distinguished by the hadronic invariant mass @xmath10 : @xmath11 for qe , and @xmath12 for res .
@xmath13 is an artificial boundary between the res and dis ( @xmath14 ) processes , and we take @xmath15 gev . figure [ region ] shows the kinematical region of each qe , res and dis process on the ( @xmath16 , @xmath17 ) plane at @xmath18 gev .
gev , width=226 ] the spin polarization vector of @xmath5 is defined as @xmath19 in the @xmath5 rest frame in which the z - axis is taken along it s momentum direction in the laboratory frame .
@xmath20 denotes the degree of polarization .
@xmath21 gives the fully polarized @xmath5 , and @xmath22 gives unpolarized @xmath5 .
the degree of polarization ( @xmath20 ) and the spin directions ( @xmath23 ) are functions of the tau energy @xmath24 and the scattering angle @xmath25 .
this spin polarization vector is related to the spin density matrix @xmath26 : @xmath27 the spin density matrix is calculated as @xmath28 , where @xmath29 is the helicity amplitude with the @xmath5 helicity @xmath30 defined in the laboratory frame , and @xmath31 is the usual spin summed cross section . the spin density matrix of @xmath5 production is obtained by using the leptonic and hadronic tensor as @xmath32 where @xmath33 is fermi constant and @xmath34 . the leptonic tensor is expressed as @xmath35 where the leptonic weak current @xmath36 is @xmath37
we show the hadronic tensors which we used in our calculation for each process , i.e. , qe , res and dis . several parameters and form factors
were updated to the recent our paper @xcite .
the hadronic tensor for the qe scattering process ; @xmath38 is written by using the hadronic weak transition current @xmath39 as follows @xcite : @xmath40 where @xmath41 is the cabibbo angle .
@xmath39 are defined as @xmath42 where @xmath43 is written in terms of the six weak form factors of the nucleon , @xmath44 , @xmath45 , @xmath46 and @xmath47 , as @xmath48\gamma_5.\end{aligned}\ ] ] we can drop @xmath49 and @xmath50 because of time reversal invariance and isospin symmetry . moreover , the vector form factors @xmath51 and @xmath52 are related to the electromagnetic form factors of nucleons under the conserved vector current hypothesis : @xmath53 where @xmath54 with a vector mass @xmath55 gev and @xmath56 . for the axial form factor @xmath45 and @xmath47
, we use : @xmath57 with @xmath58 @xcite and an axial vector mass @xmath59 gev @xcite .
notice that the pseudoscalar form factor @xmath47 plays an important role for the polarization of @xmath5 produced by neutrino because its contribution is proportional to the lepton mass and it has the spin - flip nature , although it is not known well . in ref .
@xcite , we discussed it in detail . the hadronic tensor for the @xmath0 resonance production ( res ) process ; @xmath60 is calculated in terms of the nucleon-@xmath0 weak transition current @xmath61 as follows @xcite : @xmath62 where @xmath63 is the breit - wigner factor @xmath64 with its running width @xmath65 @xmath66 gev and @xmath67
. the current @xmath61 for the process @xmath68 is defined by @xmath69 where @xmath70 is the spin-3/2 particle wave function and the vertex @xmath71 is expressed in terms of the eight weak form factors @xmath72 as @xmath73\gamma_5 { \nonumber}\\ \!\!\!\!&+&\!\!\!\ !
\frac{g_{\mu\alpha}\!\not\!q -\gamma_{\mu}q_{\alpha}}{m }
c^a_3 + \frac{g_{\mu\alpha}\,p'\cdot q -p'_{\mu}q_{\alpha}}{m^2 } c^a_4 { \nonumber}\\ \!\!\!\!&+&\!\!\!\ ! g_{\mu\alpha
} c^a_5 + \frac{q_{\mu}q_{\alpha}}{m^2 } c^a_6.\end{aligned}\ ] ] by using the isospin invariance and the wigner - eckart theorem , we obtain another nucleon-@xmath0 weak transition current as @xmath74 from the cvc hypothesis , @xmath75 and the other vector form factors @xmath76 are related to the electromagnetic form factors .
we adopt the modified dipole parameterizations @xcite : @xmath77 with @xmath78 and @xmath79 gev . for axial form factors ,
we use @xcite @xmath80 with @xmath78 and @xmath81 gev . for @xmath82 and @xmath83 , @xmath84 and @xmath85 give good agreements with the data @xcite . as in the case of the @xmath86 of the qe process
, @xmath87 has significant effects on the @xmath5 production cross section and the @xmath5 polarization @xcite . in the dis region ,
the hadronic tensor is estimated by using the quark - parton model ; @xmath88 where @xmath89 is the four - momentum of the scattering quark , @xmath90 is its momentum fraction , and @xmath91 are the parton distribution function(pdf ) s inside a nucleon . the quark tensor @xmath92 is @xmath93.\end{aligned}\ ] ] the upper sign should be taken for quarks and the lower for antiquarks .
we retain the final quark mass @xmath94 for the charm quark as @xmath95 gev , but otherwise we set @xmath96 . in the calculation , we used the mrst2002 @xcite as the pdf s ( = 1.25 gev@xmath97 in mrst2002 ) ] .
we show the spin polarization vector of @xmath5 lepton produced by neutrino off isoscalar targets as a function of its energy @xmath98 and the scattering angle @xmath25 in the laboratory frame .
figure [ taupol ] summarizes our results for the @xmath99 process at @xmath100 gev .
the top three figures show the double differential cross sections as a function of @xmath98 at the scattering angle @xmath101 ( left figures ) , @xmath102 ( center figures ) and @xmath103 ( right figures ) .
the area of the histogram for the qe process is normalized to the cross section .
the qe and res cross sections are large at forward scattering angles , and the dis contribution become more significant at large scattering angles , though the cross section gets smaller . a set of three middle figures give the degree of polarization @xmath20 , and in the bottom three figures we show the polarization direction @xmath104 in eq .
( [ vector ] ) .
the produced @xmath6 is almost fully polarized except at the very small scattering angle . as for the angle of the polarization vector ,
the high energy @xmath6 is almost left - handed ( @xmath105 ) . on the other hand ,
the spin of low energy @xmath106 turns around .
the azimuthal angle @xmath107 takes @xmath108 at all energies , which means that the spin vector points to the direction of the initial neutrino momentum axis . in order to understand the above features , it is useful to consider the polarization of @xmath6 in the center of mass ( cm ) frame of the scattering particles .
let us consider the dis process in the @xmath109 cm frame , since the @xmath109 scattering is dominant in the @xmath110 process . in this frame , produced
@xmath106 is fully left - handed polarized at all scattering angles .
this is because the initial @xmath4 and @xmath111 ( @xmath112 or @xmath113 quarks ) are both left - handed and hence angular momentum along the initial momentum direction is zero , while in the final state the produced @xmath114 quark is left - handed and hence only the left - handed @xmath6 is allowed by the angular momentum conservation .
this selection rule is violated slightly when a charm quark is produced in the final state and because of gluon radiation at higher orders of qcd perturbation theory . the @xmath6 polarization in the laboratory frame
is then obtained by the lorentz boost . in the qe and res processes ,
situations are almost the same as in the dis process . in the cm frame of @xmath115 collisions ,
the @xmath6 lepton produced by the qe or res process is almost left - handed at all angles , for the cm energy of @xmath116 gev for @xmath100 gev , for our parametrizations of the transition form factors .
high energy @xmath6 s in the laboratory frame have left - handed polarization because those @xmath6 s have forward scattering angles also in the cm frame .
however , lower energy @xmath6 s in the laboratory frame tends to have right - handed polarization because they are produced at backward angles in the cm frame .
at the zero scattering angle @xmath101 of the laboratory frame , the change in the @xmath6 momentum direction occurs suddenly , and hence the transition from the left - handed @xmath6 at high energies to the right - handed @xmath6 at low energies is discontinuous .
the information on the polarization of @xmath5 produced through the neutrino - nucleon scattering is essential to identify the @xmath5 production signal since the decay particle distributions depend crucially on the @xmath5 spin polarization .
it is needed in long baseline neutrino oscillation experiments which should verify the large @xmath117 oscillation , and is also needed for the background estimation of @xmath118 appearance experiments which should measure the small mixing angle of @xmath119-@xmath120 oscillation . in this report we presented the spin polarization of @xmath6 produced in tau - neutrino nucleon scattering via charged currents .
quasi - elastic scattering ( qe ) , @xmath0 resonance production ( res ) and deep inelastic scattering ( dis ) processes have been studied .
the three subprocesses are distinguished by the hadronic invariant mass @xmath10 .
@xmath121 gives qe , @xmath122 gives res , and @xmath14 gives dis .
here we set the kinematical boundary of res and dis process at @xmath15 gev .
the spin density matrix of @xmath5 production has been defined and the @xmath5 spin polarization vector has been defined and parametrized in the @xmath5 rest frame whose polar - axis is taken along the momentum direction of @xmath5 in the laboratory frame . the spin density matrix has been calculated for each subprocess by using the form factors for the qe and res processes , and by using the parton distribution functions for the dis process .
we have shown the spin polarizations of @xmath5 as function of the @xmath5 energy and the scattering angle in the laboratory frame for @xmath123 process at @xmath100 gev .
we found that the produced @xmath6 have high degree of polarization , but their spin directions deviate significantly from the massless limit predictions at low and moderate @xmath5 energies .
qualitative feature of the predictions have been understood by considering the helicity amplitudes in the cm frame of the scattering particles and the effects of lorentz boost from the cm frame to the laboratory frame . finally , we summarize our findings in figure [ convec ] . in figure
[ convec ] , we show the polarization vector @xmath124 of @xmath106 for the @xmath123 process at @xmath100 gev on the ( @xmath125 , @xmath126 ) plane , where @xmath127 and @xmath25 are the produced @xmath5 momentum and the scattering angle in the laboratory frame . the length of each arrow gives the degree of polarization ( @xmath128 ) at each phase - space point and its orientation gives the spin direction in the @xmath6 rest frame .
the differential cross section is described as a contour map , where only the dis cross section is plotted to avoid too much complexity .
the outer line gives the kinematical boundary , along which the qe process occurs .
figure [ convec ] is a more visual version of the information given in figure [ taupol ] . , @xmath126 ) plane for the @xmath129 process at @xmath100 gev in the laboratory frame .
the kinematical boundary is shown by the thick gray curve .
the qe process contributes along the boundary , and the res process contributes just inside of the boundary .
the @xmath6 polarization are shown by the arrows .
the length of the arrows give the degree of polarization , and the direction of arrows give that of the @xmath106 spin in the @xmath106 rest frame .
the size of the 100% polarization ( @xmath130 ) arrow is shown as a reference .
the arrows are shown along the laboratory scattering angles , @xmath101 , @xmath131 , @xmath102 , @xmath132 , and @xmath103 , as well as along the kinematical boundary .
, width=245 ]
k.m . and h.y .
are grateful to the organizers for the interesting and pleasant workshop and also for giving us the opportunity to present this work . k.m . would like to thank t. morii and m. sakuda for encouragements .
would like to thank riken bnl research center for the support of travel expense for this workshop .
9 icarus collaboration home page , http:// pcnometh4.cern.ch/. opera collaboration home page , http:// operaweb.web.cern.ch/. cngs project home page , http://proj - cngs .
k. hagiwara , k. mawatari , h. yokoya , nucl .
b 668 ( 2003 ) 364 [ arxiv : hep - ph/0305324 ] . k.s .
kuzmin , v.v .
lyubushkin , v.a .
naumov , hep - ph/0312107 .
they considered tau polarization with a covariant method .
llewellyn smith , phys .
rep . 3 ( 1972 ) 261 .
s. eidelman et al .
( particle data group ) , phys .
b 592 ( 2004 ) 1 .
v. bernard , l. elouadrhiri , u. meissner , j. phys .
g 28 ( 2002 ) r1 .
k. hagiwara , k. mawatari , h. yokoya , phys .
b 591 ( 2004 ) 113 [ arxiv : hep - ph/0403076 ] . p.a . schreiner and f. von hippel , nucl .
b 58 ( 1973 ) 333 . s.k .
singh , m.j .
vicente vacas , e. oset , phys .
b 416 ( 1998 ) 23 . e.a .
paschos , j.y .
yu , m. sakuda , phys .
d 69 ( 2004 ) 014013 .
olsson , e.t .
osypowski , e.h .
monsay , phys .
d 17 ( 1978 ) 2938 . | we studied polarization effects in tau production by neutrino - nucleon scattering . quasi - elastic scattering , @xmath0 resonance production and deep inelastic scattering processes are taken into account for the cern - to - gran sasso projects .
we show that the tau produced by neutrino has high degree of polarization , and its spin direction depends non - trivially on the energy and the scattering angle of tau in the laboratory frame . |
the investigation of multi - spin interactions has been initiated several decades ago in order to clarify their influence on phase transitions and magnetic properties in various physical systems . in order to investigate basic aspects of multi - spin interactions
the authors have utilized various theoretical methods including exact calculations @xcite-@xcite , series expansions @xcite-@xcite , renormalization - group techniques @xcite-@xcite , monte carlo simulations @xcite-@xcite , mean - field and effective - field theory @xcite-@xcite . on the other hand , from the experimental point of view , the models with multi - spin interactions have been widely used to explain the thermodynamic properties of various physical systems , such as binary alloys @xcite , classical fluids @xcite , solid he@xmath0 @xcite , lipid bilayers @xcite , metamagnets @xcite , rare gases @xcite or hydrogen bonded ferroelectrics pbhpo@xmath1 and pbdpo@xmath1 @xcite .
one should also mention here that the models with multi - spin interactions have been successfully used to describe the first - order phase transition in the squaric acid crystal h@xmath2c@xmath2o@xmath1 @xcite and some co - polymers @xcite .
moreover , the cycling four - spin exchange interactions have been adopted to explain experimental results on spin gaps @xcite-@xcite , raman peaks @xcite and optical conductivity of the cuprate ladder la@xmath2ca@xmath3cu@xmath4o@xmath5 @xcite .
the four - spin interactions have been taken into account also in the study of two - dimensional antiferromagnet la@xmath2cu0@xmath1 , the parent material of high - t@xmath6 superconductors @xcite .
a special kind of higher - order spin interactions that are known as three - site four - spin ones has been introdudes an widely studied iwashita and uryu @xcite-@xcite .
finally , let us mention a series of works by kbler et al .
@xcite-@xcite in which the authors have very carefully investigated the role of higher - order spin interactions in a wide class of real magnetic materials .
as far as it concerns of magnetic properties , the models with multi - spin interactions may exhibit some peculiarities , for example , the non - universal critical behavior @xcite or deviations from the bloch s t@xmath7 law at low temperatures @xcite-@xcite .
here it is worth emphasizing that some of these phenomena are not yet well understood and clarified even at the present time .
in fact , the investigation of many - body interactions is of tremendous importance in all branches of physics , since such studies may discover new physical phenomena that do not appear in the systems with pair interactions only .
however , it is necessary to recall that the investigation of the systems with many - body interactions is as a rule much more complex than those with pair interactions only .
nonetheless , we have recently demonstrated @xcite-@xcite that various versions of the ising model represent a very good theoretical ground for an accurate treatment of multi - spin interactions .
the main aim of this work is to extend our recent research in this field in order to investigate in detail the role of so - called three - site four - spin interactions in crystalline systems with localized magnetic moments . for this purpose
, we will study the mixed - spin 1/2 and 1 ising model with the single - ion anisotropy , pair and three - site four spin interactions on a decorated square lattice .
the outline of the present work is as follows . in sec .
[ theory ] and [ gstp ] we derive exact equations for all physical quantities applying a generalized form of decoration - iteration transformation .
the ground - state and finite - temperature phase diagrams are discussed in detail in sec .
[ nr ] along with thermal variations of other physical quantities .
finally some conclusions are sketched in the last section .
in this work we will investigate the mixed spin-1/2 and spin-1 ising model on a decorated square lattice depicted in fig .
[ fig1 ] .
the system is described by the hamiltonian @xmath8 where @xmath9 and @xmath10 respectively denote the nearest - neighbor and next - nearest - neighbor bilinear exchange interactions , @xmath11 is the three - site four - spin exchange interaction and @xmath12 represents the single - ion anisotropy parameter .
the summations in the first and second term in ( [ eq1 ] ) run over all relevant pairs on the decorated square lattice , while in the third and fourth term the summations are over all decorating spins , i.e. @xmath13 , where @xmath14 represents the total number of spin-1/2 atoms . in order to apply the decoration - iteration transformation to the present model ,
we at first express the total hamiltonian in the form @xmath15 where the hamiltonian @xmath16 includes all interaction terms within the _ k_-th bond of the lattice and it is given by @xmath17 now , using ( [ eq2 ] ) , the partition function of the model can be expressed as @xmath18 where the curled brackets denote the fact that relevant summation concerns all spin variables of the lattice .
introducing the following decoration - iteration transformation @xcite @xmath19 one rewrites equation ( [ eq4 ] ) as follows @xmath20 in the last equation @xmath21 represents the partition function of conventional ising model on a square lattice described by the hamiltonian @xmath22 .
this partition function has been exactly calculated in a seminal onsager s work @xcite and it will be used in this paper to obtain exact results for thermodynamic properties of the model under investigation . of course , to complete the calculation of @xmath23 we have also to determine the unknown functions @xmath24 and @xmath25 . fortunately , this evaluation may be straightforwardly performed by substituting @xmath26 and @xmath27 into eq .
( [ eq5 ] ) and in this way one gets @xmath28 where @xmath29 @xmath30
the ground - state phase diagram can be determined investigating the internal energy of the system at @xmath31 .
since we do not consider any external field , the internal energy of the system can be evaluated as a mean value of the hamiltonian ( [ eq1 ] ) , i.e. @xmath32 and it takes the following form @xmath33 where the angular brackets denote the standard canonical averaging using the density matrix @xmath34 . for further progress in calculation
it is of crucial importance that all correlation functions entering previous equation can be calculated using the generalized callen - suzuki identities @xcite which in our case take the form @xmath35 } { 2 \mbox{e}^{\beta j_4 \mu_{ki}^z \mu_{kj}^z } \cosh \bigl [ \beta j ( \mu_{ki}^z + \mu_{kj}^z ) \bigr ] + \mbox{e}^{-\beta d } } \biggr \rangle,\ ] ] @xmath36 } { 2 \mbox{e}^{\beta j_4 \mu_{ki}^z \mu_{kj}^z } \cosh \bigl [ \beta j ( \mu_{ki}^z + \mu_{kj}^z ) \bigr ] + \mbox{e}^{-\beta d } } \biggr \rangle,\ ] ] where @xmath37 represents a function of arbitrary spin variables except of the variable @xmath38 . now applying the well - known differential operator technique @xcite , we recast previous equation in the more convenient form @xmath39 and @xmath40 where the coefficients @xmath41 are listed in the appendix .
it is clear that after substituting @xmath42 into eq .
( [ eq13 ] ) and eq .
( [ eq14 ] ) , we respectively obtain equations for the evaluation of sublattice magnetization @xmath43 and quadrupolar moment @xmath44 . similarly , setting @xmath45
one gets expressions for the correlation functions @xmath46 and @xmath47 , that are necessary for evaluation of the internal energy according to ( [ eq10 ] ) .
finally , it also very important to notice that the magnetization of nodal spins @xmath48 and also the correlation function @xmath49 appearing on the r.h.s of ( [ eq13 ] ) and([eq14 ] ) can be simply evaluated from the relations @xmath50 and @xmath51 where the @xmath52 represents relevant canonical averaging on the original ( non - decorated ) square lattice using the density matrix @xmath53 .
calculation of other thermodynamic quantities is a straightforward process , since in the previous section we have obtained an exact relation for the partition function of the model under investigation ( see eqs ( [ eq6])-([eq9 ] ) ) . at first , the helmholtz free energy is easily obtained in the form @xmath54 where parameters @xmath55 , @xmath56 are given by eq .
( [ eq7 ] ) and @xmath57 represents the helmholtz free energy of the ising square lattice @xcite @xmath58d\phi\ ] ] with @xmath59 it is clear that the contribution from decorating spins to the total helmholtz free energy is represented by the first term in eq .
( [ eq17 ] ) which is an analytic function in the whole parameter space .
therefore the critical behavior of the model will necessarily belong to the same universality class as that one of the usual 2d ising model .
consequently , the finite - temperature phase boundaries of the system under investigation can be simply determined by substituting the inverse critical temperature of the square lattice @xmath60 into eq .
( [ eq7 ] ) , i.e. @xmath61 with @xmath62 . in this work
we will also study the entropy and specific heat that are respectively given by the relations @xmath63 and @xmath64 here the subscript zero indicates the fact that relevant quantity is calculated in the zero external magnetic field .
numerical results obtained from the above equations are discussed in detail in the next section .
in this section we will discuss the most interesting numerical results for the ground - state and finite - temperature phase diagrams , and thermal dependencies of magnetization , correlation functions , internal energy , entropy and specific heat .
for this purpose it is useful to introduce the following dimensionless parameters : @xmath65 , @xmath66 and @xmath67 . in order to investigate the physical nature of possible phases at the ground state
, we have numerically studied the sublattice magnetizations , quadrupolar moment and various correlations functions in our system at @xmath31 . on the basis of these calculations ,
the ground - state phase boundaries have been established from the minimum values of the internal energy of the system .
our results are summarized in fig .
[ fig2 ] , where we have depicted the ground - state phase diagram in the @xmath68 space for the special case of @xmath69 , i.e. @xmath70 . as one can see from the figure , the ordered ferrimagnetic phase becomes stable in the region above the line @xmath71 for negative values of nearest - neighbor pair exchange interaction ( @xmath72 ) . on the other hand , for @xmath73 the ferromagnetic ordering appears in the region above the line @xmath74 .
it also clear that for @xmath75 the ordered ferrimagnetic or ferromagnetic phase is respectively stable for arbitrary negative or positive nonzero values of @xmath76 .
space for the mixed - spin ising model without nex - nearest pair interaction ( @xmath69 ) . ]
next , it follows from our analysis that a disordered paramagnetic phase becomes stable in the ground state for arbitrary values of @xmath76 whenever the crystal - field parameter @xmath77 satisfies inequality @xmath78 , i.e. when the single - ion anisotropy takes negative and strong enough values .
consequently , the ordered phases coexist with the paramagnetic one along the line @xmath79 and the system exhibits a first - order phase transition when crossing this line .
now , let us look closely at the particular case of system with pure three - site interaction which is obtained by setting @xmath80
. in this case one finds from eq .
( [ eq13 ] ) that the sublattice magnetization @xmath81 becomes zero for arbitrary values of @xmath77 and of course , the same statement applies also for all correlations including a spin of @xmath82 sublattice .
moreover , one finds from eq .
( [ eq14 ] ) that @xmath83 for @xmath84 , thus the spin states @xmath85 and @xmath86 of all decorating atoms are equally likely occupied at @xmath87 , while the sublattice @xmath24 exhibits the long - range order with @xmath88 . thus the system as a whole exhibits an unexpected behavior with the non - zero ground - state entropy @xmath89 belonging to the partially ordered magnetic phase .
this surprising behavior appears as a result of the interplay between three - site four - spin interaction @xmath11 and crystal - field parameter @xmath12 . on the other hand , for @xmath90 and @xmath91
we obtain from eq .
( [ eq14 ] ) @xmath92 indicating that all decorating atoms occupy the spin states @xmath93 , so that each atom on the @xmath24 sublattice is surrounded by decorating atoms with @xmath93 . as a result of this spin configuration ,
each atom of the @xmath24 subllatice can be found equally likely in the spin state @xmath94 or @xmath95 thus the entropy at @xmath87 takes now the value @xmath96 . in this case , of course , no magnetic order appears in the ground state .
moreover , one should emphasize here that at the point with co - ordinates @xmath97 and @xmath98 one observes a very interesting partially ordered phase with @xmath99 , @xmath100 , nonzero @xmath101 and unusually high entropy @xmath102 at @xmath87 . as far as we know , such a phase has not been reported in the literature on the ising model until now . in order to complete the ground - state analysis , let us briefly mention the effect of a positive next - nearest - neighbor interaction , i.e. @xmath103 .
it is clear that in this case the disordered paramagnetic phase never becomes stable for @xmath104 at @xmath31 , so that the system will always exhibit the long - range order with @xmath105 for @xmath106 , and @xmath107 for @xmath108 . on the other hand ,
the situation for @xmath97 , @xmath109 and @xmath103 becomes identical with the above discussed case with @xmath97 , @xmath109 and @xmath69 .
finally , for @xmath97 , @xmath110 and @xmath103 , the system will exhibit a long - range order on the @xmath24 sublattice , despite of the fact that all atoms of the @xmath82 sublattice will occupy the zero - spin states .
however , contrary to the case with @xmath69 , the entropy of the system now always vanishes in the limit of @xmath111 .
now , let us proceed with the discussion of finite - temperature phase diagrams and compensation temperatures .
we recall that the critical temperature is calculated from eq .
( [ eq20 ] ) and takes the same value for both ferromagnetic and ferrimagnetic cases . in ferrimagnetic systems ,
the compensation temperature is defined as a temperature at which the total magnetization vanishes below the critical temperature , so that in our case it is determined by the condition @xmath112 . in what follows
, we will discuss the most interesting results of phase diagrams and compensation temperatures obtained numerically for some characteristic combinations of freely adjustable parameters . at first , in fig .
[ fig3a ] there are depicted the dependencies of critical temperatures on the reduced single - ion anisotropy parameter @xmath77 for @xmath113 and 1 . as we can see from the figure ,
the phase boundaries for arbitrary @xmath104 do not depend on the sign of the nearest - neighbor interaction , so that both the ferromagnetic @xmath114 and ferrimagnetic @xmath115 phases always vanish at the same critical temperature .
this statement follows , of course , directly form eq .
( [ eq20 ] ) which is invariant under changing the sign of @xmath9 . in agreement with our discussion in previous section
, one finds that magnetization of the system for both ordered phases takes its saturation value at @xmath116 in the region of @xmath117 . of course
, by increasing the temperature , the magnetization will gradually decrease until it continuously vanishes at some critical temperature @xmath118 above which the disordered paramagnetic phase becomes stable .
moreover , as we have already mentioned above , the relevant second - order phase transitions belong to the 2d ising universality class .
next , one also observes from fig .
[ fig3a ] that the region of stability of paramagnetic phase extends from high temperatures down to zero absolute temperature for arbitrary values of parameter @xmath76 , whenever @xmath119 . here
one should also notice that if @xmath69 then no compensation points appear in the system , regardless of the values of parameters @xmath76 and @xmath77 .
.5 space for @xmath69 and for several typical values of @xmath76 .
the inset shows the detail view for @xmath120 where the reentrant behavior with two critical temperatures appears .
+ ( b ) - the same as in the case ( a ) but for @xmath121 .
the critical boundaries are depicted by full lines and compensation temperatures by dashed ones .
the inset shows the existence of two compensation temperatures in a narrow region of negative values of @xmath77.,title="fig : " ] .5 space for @xmath69 and for several typical values of @xmath76 .
the inset shows the detail view for @xmath120 where the reentrant behavior with two critical temperatures appears .
+ ( b ) - the same as in the case ( a ) but for @xmath121 .
the critical boundaries are depicted by full lines and compensation temperatures by dashed ones .
the inset shows the existence of two compensation temperatures in a narrow region of negative values of @xmath77.,title="fig : " ] now , let us inspect the special case of the system with the pure three - site four - spin interaction , i.e. @xmath97 and @xmath69 .
in this case one find that for @xmath109 , the partially ordered phase with @xmath122 and @xmath123 is stable bellow @xmath118 , while above the critical temperature becomes stable again the standard paramagnetic phase .
moreover , looking carefully on the phase boundary one finds the c - shaped form of this curve indicating the appearance of reentrant behavior with two phase transitions in a narrow low - temperature region in the neighborhood of @xmath124 ( see the inset in fig .
[ fig3a ] ) .
space for @xmath97 and for several typical values of @xmath77 .
the inset shows the detail view for @xmath125 where the reentrant behavior with three critical temperatures appears . ]
as we have already mentioned earlier , the non - zero next - nearest - neighbor pair interaction ( @xmath126 ) eliminates the existence of paramagnetic phase at @xmath127 , regardless of the values of other parameters , so that the phase boundaries must significantly change . for @xmath121 and for the same values of @xmath76 as in fig .
[ fig3a ] , the situation is illustrated in fig .
[ fig3b ] where we have depicted using full lines some typical phase boundaries .
.5 space for @xmath69 ( a ) and @xmath128 ( b ) , that is without and with the small next - nearest - neighbor interaction , respectively.,title="fig : " ] .5 space for @xmath69 ( a ) and @xmath128 ( b ) , that is without and with the small next - nearest - neighbor interaction , respectively.,title="fig : " ] regarding the compensation temperatures one should stress here that , in general , the compensation effect is only possible for relatively strong nonzero values of @xmath129 , which keeps the sublattice magnetization @xmath130 strong enough over a large temperature region .
the typical dependencies of the compensation temperature on the parameter @xmath77 are shown in fig .
[ fig3b ] by dashed lines for @xmath121 , @xmath131 and @xmath132 .
the inset in this figure illustrates that even two compensation temperatures are possible in a narrow region of negative values of @xmath77 .
as far as we know , such a finding has not been reported yet for the systems with higher order interactions .
moreover , a closer investigation of the particular case with @xmath133 and @xmath126 indicates that the re - entrant behavior with three different critical temperatures appears in the neighborhood of @xmath134 .
this phenomenon is most clearly visible when the phase diagrams are constructed in the @xmath135 space as it is shown in fig .
[ fig4 ] . finally , to in order to obtain a global view of the critical surface , we have combined together different two dimensional phase diagrams and we have obtained the three - dimensional global phase diagrams of the system that are presented in figs . [ fig5a ] and [ fig5b ] for @xmath69 and @xmath128 , respectively . in this part
we present the most interesting numerical results for thermal dependencies of thermodynamic quantities . in order to demonstrate the role of the three - site four - spin interaction , we at first in figs .
[ fig6a ] and [ fig6b ] show the temperature dependencies of sublattice magnetization and quadrupolar moment @xmath136 for @xmath137 and for several representative values of @xmath77 .
let us recall that the sublattice @xmath82 does not exhibit the long - range order , since the magnetization @xmath81 vanishes for arbitrary values of @xmath77 and at any temperature , whenever @xmath97 holds .
.5 and quadrupolar momentum @xmath138 of the decorated mixed - spin ising system for @xmath139 and for several typical values of @xmath77.,title="fig : " ] .5 and quadrupolar momentum @xmath138 of the decorated mixed - spin ising system for @xmath139 and for several typical values of @xmath77.,title="fig : " ] consequently , at @xmath87 the system as a whole will be partially ferromagnetically ordered for @xmath140 with @xmath141 and @xmath142 as it is clearly illustrated in figs .
[ fig6a ] and [ fig6b ] for @xmath143 and @xmath144 .
moreover , at the ground state for @xmath137 and @xmath145 one finds that the sublattice magnetization takes the value @xmath146 and the quadrupolar moment @xmath147 , instead of their saturation values 0.5 and 1 , respectively .
these values indicate that both sublattices exhibit very interesting frustrated behavior which originates from the effect of the three - site four - spin interaction and , as far as we know , such a finding has not been reported for the ising models until now . .
the parameters of the model are fixed to the selected specific values in order to illustrate the cases with one and two compensation temperatures . ] finally , selecting the value of @xmath77 slightly bellow @xmath148 one observes a reentrant behavior in the temperature behavior of the magnetization @xmath130 as it is shown in fig .
[ fig6a ] for @xmath149 .
of course , one can very simply verify that all thermal dependencies are in a perfect agreement with the ground - state and finite - temperature phase diagrams discussed in previous section , since @xmath150 at @xmath31 . in order to complete our analysis of the magnetization ,
let us recall that for non - zero values of @xmath129 , the most interesting thermal variations of the magnetization appears in the ferrimagnetic case ( i.e , @xmath108 ) in the region where compensation effects take place . to illustrate this original behavior ,
we have depicted in fig .
[ fig7 ] the magnetization curves exhibiting one ( the red curve ) and two compensation temperatures ( the green curve ) .
.5 and for several typical values of @xmath77 .
+ ( b ) - thermal dependencies of reduced specific heat of the decorated mixed - spin ising system for @xmath151 and the same values of @xmath77 as in case ( a).,title="fig : " ] .5 and for several typical values of @xmath77 .
+ ( b ) - thermal dependencies of reduced specific heat of the decorated mixed - spin ising system for @xmath151 and the same values of @xmath77 as in case ( a).,title="fig : " ] it follows from the previous discussion that the system with pure three - site four - spin interaction will necessarily exhibit finite values of entropy in the ground state .
this interesting phenomenon is investigated in fig .
[ fig8a ] , where we have depicted the temperature dependencies of entropy for @xmath137 and several values of @xmath77 corresponding to figs .
[ fig6a ] and [ fig6b ] .
the results shown in this figure prove that the ground - state entropy of the system may take only three different values depending on the value of @xmath77 , namely , @xmath152 here one should notice that the value of @xmath153 originates solely from the contribution of the @xmath24 sublattice , while the second non zero value of @xmath154 represents exclusive contribution of the @xmath82 sublattice . of course , these two values correspond to ground - state spin configurations described in detail in the subsection _
[ gs]_. on the other hand , the value of entropy for @xmath145 consists apparently from two parts , i.e. , @xmath155 , where the major contribution comes from the @xmath82 sublattice where all three spin states @xmath156 are randomly occupied at @xmath31 ( see green curve in fig .
[ fig8a ] ) . the small supplementary value @xmath157 represents the contribution of the @xmath24 sublattice which is also slightly disordered at @xmath31 since the sublattice magnetization @xmath130 does not reach its saturated value ( see green curve in fig . [ fig6a ] ) .
here we should also mention that the results for ground state entropy in case of @xmath158 , that is when next - nearest interaction is present , revealed just two values of entropy at @xmath87 , from which one of them equals 0 ( see also ground state analysis in subsection [ gs ] ) . comparing different entropy behavior in ground states for pure three - site four - spin interaction and those with additional billinear interactions ( @xmath9 , @xmath10 )
one can see a significant signature of three - site four - spin interaction in pushing the system to the unconventional partially ordered states .
finally , we have calculated the temperature dependencies of the magnetic specific heat and the results obtained for a special case of @xmath159 are presented in fig .
[ fig8b ] . as one can see from the figure , the curves for @xmath160 and @xmath161 exhibit at critical point a logarithmic singularity similarly as the usual spin-1/2 ising model on a square lattice .
similarly , for strong negative values of @xmath77 one observes the expected behavior for paramagnetic systems . on
should emphasize here that despite of non - zero ground - state entropy ( cf .
[ fig8a ] ) , the specific heat goes always to zero for @xmath162 , so that the third law of thermodynamics is perfectly satisfied .
in this work we have concentrated on clarifying the influence of three - site four - spin interactions on magnetic properties of a mixed - spin ising model on the decorated square lattice including also the pair exchange interaction and crystal - field contributions . applying the generalized decoration - iteration transformation ,
we have exactly obtained all relevant physical quantities of the model , including the ground - state and finite - temperature phase diagrams .
the numerical results obtained in this work clearly illustrate the principal influence of higher - order spin interactions on all relevant physical quantities .
the most original behavior has been observed in the case with pure three - site four spin interactions , where we have confirmed that unlike of four - site four - spin interactions ( see @xcite-@xcite ) the three - site four - spin interactions may initiate the appearance of a partial long - range order .
moreover , due to the very strong frustrations in the system one observes a non - zero entropy at @xmath87 over a wide range of parameters . in general ,
it is necessary to emphasize that the multi - spin interactions have as a rule very different symmetries in comparison with those of the pair ( bilinear ) interaction terms and , of course , these new symmetries basically determine the behavior of the system .
it is well known that the theoretical investigation of the systems with many - body interactions is extraordinarily complicated task , thus the absolute majority of the papers in diverse research fields treat only the systems with two - body ( pair - wise ) interactions .
one should , however , notice that various versions of ising and heisenberg models represent rare exceptions that straightforwardly enable to account for many different forms of multi - spin ( i.e. many - body ) interactions .
for that reason the localized - spin models represent an excellent basis for deep understanding of various many - body interactions going beyond the standard pair - wise picture .
we hope that the present study will may initiate a wider interest in investigation of magnetic systems with multi - spin interactions .
+ the coefficients @xmath163 , @xmath164 , @xmath165 are respectively given by @xmath166 \\
\nonumber & & a_1 = f(j , \frac{j}{4})\\ \nonumber & & a_2 = \frac{1}{2}\biggl [ g\bigl(j , \frac{j_4}{4}\bigr ) - g\bigl(0 , -\frac{j_4}{4 } \bigr ) \biggr ] \end{aligned}\ ] ] where @xmath167 90 f. y. wu , phys . rev .
b 4 ( 1971 ) 2312 .
l. p. kadanoff and f. j. wegner , phys .
b 4 ( 1971 ) 3989 .
f. y. wu , phys .
a 38 ( 1972 ) 77 .
r. v. ditzian , phys .
a 42 ( 1972)67 .
m. suzuki , phys .
( 1972 ) 507 .
t. horiguchi , physica a 130 ( 1985 ) 194 .
t. horiguchi , l.l .
goncalves , physica a 133 ( 1985 ) 460 .
t. morita , t. horiguchi , physica a 136 ( 1986 ) 99 .
f. wang , m. suzuki , physica a 230 ( 1996 ) 639 .
a. lipowski , physica a 248 ( 1998 ) 207 .
s. lackov , m. jaur , phys .
e 64 ( 2001 ) 036 126 .
s. lackov , m. jaur , czech . j. phys
. 52 ( 2001 ) a33 .
s. lackov , m. jaur , physica a 326 ( 2003 ) 189 .
s. lackov , m. jaur , t. horiguchi , physica a 339 ( 2004 ) 416 .
m. jaur , v. tuba , k. szalowski , t. balcerzak , acta phys .
polonica 126 ( 2014 ) 48 .
j. oitmaa , r.w .
gibberd , j. phys .
c 6 ( 1973 ) 2077 .
griffiths , d.w .
wood , j. phys .
c 7 ( 1974 ) l54 .
griffths , d.w .
wood , j. phys .
c 7 ( 1974 ) 4021 .
m. nauenberg , b. nienhuis , phys .
33 ( 1974 ) 944 .
van leeuwen , phys .
34 ( 1975 ) 1056 .
nightingale , phys .
lett . 59a ( 1977 )
468- f. lee , h.h .
chen , f.y .
wu , phys .
b 40 ( 1989 ) 4871 . d.f .
styer , m.k .
phani , j.l .
lebowitz , phys .
b 34 ( 1986 ) 3361 . k.k .
chin , d.p .
landau , phys .
b 36 ( 1987 ) 275 .
heringa , h.w.j .
blwote , a. hoogland , phys .
63 ( 1989 ) 1546 .
zhang , c .- z .
yang , phys .
b 48 ( 1993 ) 9487 . c.l .
wang , z.k .
qin , d.l .
lin , j. magn .
88 ( 1990 ) 87 . t. kaneyoshi , t. aoyama , j. magn . magn .
96 ( 1991 ) 67 .
chakraborty , j. magn .
114 ( 1992 ) 155 .
s. lackov , m. jaur , j. magn .
217 ( 2000 ) 216 .
m. grimsditch , p. loubeyre and a. polian , phys .
b 33 ( 1986 ) 7192 .
m. roger , j. h. hetherington and j. m. delrieu , rev .
55 ( 1983 ) 1 .
h. l. scott , phys .
rev . a * 37 * ( 1988 ) 263 .
z. onyszkiewicz , phys .
68 a ( 1978 ) 113 . j. a. barker , phys . rev .
57 ( 1986 ) 230 .
w. chunlei , q. zikai and z. jingbo , ferroelectrics 77 ( 1988 ) 21 .
wang , z.k .
qin , d.l .
lin , phys .
b 40 ( 1980 ) 680 . c.l .
wang , z.k .
qin , d.l .
lin , solid state commun .
71 ( 1989 ) 45 .
silva , b.v .
costa , r.l .
moreira , polymer 34 ( 1993 ) 3107 .
s. brehmer , h .- j .
mikeska , m. mxuller , phys .
b 60 ( 1999 ) 329 .
m. matsuda , k. katsumata , phys .
b 62 ( 2000 ) 8903
. m. matsuda , k. katsumata , r.s .
eccleston , s. brehmer , h .- j .
mikeska , j. appl .
87 ( 2000 ) 6271 .
schmidt , c. knetter , g.s .
uhrig , europhys .
56 ( 2001 ) 877 .
nunner , p. brune , t. kopp , m. windt , m. grxuninger , acta phys . pol .
b 34 ( 2003 ) 1545 .
y. honda , y. kuramoto , t. watanabe , phys .
b 47 ( 1993 ) 11329 .
r. coldea , s.m .
hayden , g. aeppli , t.g .
perring , c.d .
frost , t.e .
mason , s .- w .
cheong , z. fisk , phys .
86 ( 2001 ) 5377 . t. iwashita and n. uryu , j. phys .
c : solid state phys .
17 ( 1984 ) 855 .
t. iwashita and n. uryu , j. phys .
c : solid state phys . 21 ( 1988 ) 4783 .
t. iwashita and n. uryu , phys .
lett . a 125 ( 1987 ) 5 .
t. iwashita and n. uryu , phys .
lett . a 155 ( 1991 ) 241 .
u. kbler , r. mueller , l. smardz , d. maier , k. fischer , b. olefs , w. zinn , z. phys .
b 100 ( 1996 ) 497 . e. mller - hartmann , u. kbler , l. smardz , j. magn .
173 ( 1997 ) 133 .
u. kbler , r.m .
mueller , w. schnelle , k. fischer , j. magn .
188 ( 1998 ) 333 .
u. kbler , a. hoser , m. kawakami , t. chatterji , j. rebizant , j. magn .
205 ( 1999 ) 343 .
i. syozi , prog .
( 1952 ) 341 .
i. syozi , in _ phase transition and critical phenomena _ , edited by c. domb and m.s .
green academic press , new york , 1972 .
m. fisher , phys .
( 1959 ) 969 l. onsager , phys.rev .
65 ( 1944 ) 117 . h. b. callen , phys . lett .
4 ( 1963 ) 161 .
m. suzuki , phys .
( 1965 ) 267 .
r. honmura and t. kaneyoshi , j. phys .
c 12 ( 1979 ) 3979 . | competitive effects of so - called three - site four - spin interactions , single ion anisotropy and bilinear interactions is studied in the mixed spin-1/2 and spin-1 ising model on a decorated square lattice .
exploring the decoration - iteration transformation , we have obtained exact closed - form expressions for the partition function and other thermodynamic quantities of the model . from these relations , we have numerically determined ground - state and finite - temperature phase diagrams of the system .
we have also investigated temperature variations of the correlation functions , internal energy , entropy , specific heat and helmholtz free energy of the system . from the physical point of view
, the most interesting result represents our observation of a partially ordered ferromagnetic or phase in the system with zero bilinear interactions .
it is remarkable , that due to strong frustrations disordered spins survive in the system even at zero temperature , so that the ground state of the system becomes macroscopically degenerate with non - zero entropy .
introduction of arbitrarily small bilinear interaction completely removes degeneracy and the entropy always goes to zero at the the ground state .
frustration , ising model , many - body interactions , exact results , decorated lattice , phase transitions . |
the interplay between superconductivity and quantum confinement has attracted growing interest due to the unique phenomena which appear if the electron motion is limited to the size smaller than the coherence length.@xcite the huge progress in nanotechnology which has been made in the last decade allows to prepare uniform ultrathin films @xcite with new superconducting properties which were predicted many years ago.@xcite the oscillations of the superconducting energy gap as a function of the nanofilm thickness were predicted by blatt and thompson@xcite in 1963 .
the authors of ref . showed that the quantum - well states which are created due to the confinement of the electron motion in the direction perpendicular to the film , greatly modulate the electron density of states near the fermi level and thus significantly affect the superconducting energy gap of the nanofilm . as a result
, one would expect a significant increase of the energy gap each time the bottom of a subband passes through the fermi sphere . nevertheless , either the experimental studies carried out in these years did not exhibit such an effect or the observed oscillations differed quantitatively from the theoretical predictions.@xcite this inconsistency was attributed to the technological difficulties in the preparation of uniform films which were typically polycrystalline and contained a large number of defects . since then , many technological obstacles were overcame what reopened the issue for the quantum confinement effect on the superconductivity in the nanoscale regime .
recently , guo at al.@xcite have fabricated ultrathin pb films on si(111 ) substrate and observed the oscillations of the critical temperature as a function of the number of atomic monolayers .
the origin of @xmath1 oscillations associated with the quantum confinement has been confirmed independently by the measurements of the quantum - well energies by using photoemission spectroscopy .
the results presented in ref . have concerned only films with the thickness greater than @xmath2 ml because of problems with stability of the even - layered films below @xmath3 ml .
the study of pb nanofilms has been extended by eom et al.@xcite who have reported @xmath1 oscillations in the epitaxially grown crystalline pb films in the thickness range @xmath4 ml . in ref .
the critical temperature has been measured using scanning tunneling micoroscopy which allows to avoid ambiguities associated with the au layer which has to be applied in the transport measurements reported in ref . .
it has been found that there is a direct correlation between the oscillations of the density of states at the fermi level and @xmath1 oscillations . also , the measurements of the critical temperature for pb nanofilms on si(111 ) substrate have revealed one more interesting phenomena called bilayer or even - odd oscillations .
as it has shown , the critical temperature for nanofilms with even number of monolayers is greater than the one for nanofilms with odd number of monolayers in proper ranges of nanofilm thickness .
this dependence reverses ( odd monolayer films have the critical temperature greater than even monalayer films ) with the periodicity of @xmath5 ml depending on the experiment.@xcite this effect has been theoretically investigated by shanenko et al . in ref . and observed in many experiments.@xcite the new direction of studies devoted to superconductivity in nanoscale regime concerns the effect of quantum confinement on the superconductor to normal metal transition induced by the magnetic field .
it has been theoretically shown , that a cascade of jumps in the energy gap as a function of the magnetic field occurs in the metallic nanowires.@xcite this effect has been explained in terms of depairing in the subsequent bands when the magnetic field increases .
@xcite the oscillations of the perpendicular upper critical field in ultrathin lead films have been reported by bao et al . in ref . .
this experimental report has shown the even - odd oscillations of the upper critical field for pb nanofilms .
it is also worth mentioning that very recently an interesting effect has been found in amorphous pb nanofilm .
@xcite namely , the increase of the critical temperature has been observed after application of a parallel magnetic field to the sample . according to our knowledge , the theoretical study of the superconducting - normal metal transition induced by the in - plane magnetic field and its interplay with the quantum confinement has not been reported until now . in the present paper we investigate the superconductor - normal metal transition driven by the in - plane magnetic field for pb nanofilms .
we show that the critical field oscillates as a function of the nanofilm thickness with the period of @xmath0 ml ( even - odd oscillations ) modulated by the beating effect with the periodicity of @xmath6 ml .
the calculated paramagnetic critical field differs from the clogston - chandrasekhar paramagnetic bulk limit.@xcite we also discuss the thermal effect and determine formulas for @xmath7 in the limit @xmath8 . finally , we take into account the oscillatory behavior of the electron - phonon coupling occurring in nanofilms and calculate @xmath9 as a function of the nanofilm thickness for the experimental samples for which a good agreement of theory and experiment has been found in the study of the critical temperature .
@xcite the present paper is organized as follows . in sec .
[ sec : model ] we introduce the basic concepts of the calculation scheme based on the bogoliubov - de gennes equations . in sec .
[ sec : results ] we analyze the results , while the summary is included in sec .
[ sec : concl ] .
the superconducting properties of the conventional phonon - mediated pairing system for the high quality single crystalline structures can be described with the use of the bcs theory , which leads to the bogoliubov - de gennes ( bdg ) equations in the form @xmath10 where @xmath11 and @xmath12 are the spin - dependent electron - like and hole - like wave functions ( @xmath13 ) , @xmath14 is the superconducting energy gap and @xmath15 is the quasi - particle energy . in the presence of the in - plane magnetic field
@xmath16 the hamiltonian @xmath17 is given by @xmath18 where @xmath19 corresponds to the spin @xmath20 and @xmath21 is related to the spin @xmath22 , @xmath23 is the free electron mass , @xmath24 is the bohr magneton , @xmath25 is the chemical potential and @xmath26 is the vector potential related to the parallel magnetic field . + in the present paper we neglect the orbital effects and consider the superconducting nanofilms in the clean paramagnetic limit .
this approximation is justified for the nanofilms with the thickness less than the magnetic length @xmath27 .
the use of the clogston - chandrasekhar paramagnetic field for pb @xmath28 t gives @xmath29 nm which means that the paramagnetic approximation can be used for pb nanofilms with thickness less than @xmath30 ml ( we assume the lattice constant for pb , @xmath31 nm ) . in the paramagnetic limit the hamiltonian can be further simplified to the form @xmath32 if we assume the periodic boundary conditions in the @xmath33 plane , the quasi - particle wave functions can be expressed as @xmath34 where @xmath35 denotes the opposite spin . in the equation ( [ wv ] )
the index @xmath36 has been replaced by @xmath37 , where @xmath38 , @xmath39 are the free electron wave vector components in the @xmath40 and @xmath41 direction , while @xmath42 labels the quantum states in the @xmath43 direction . by substituting the wave function given by ( [ wv ] ) into the bdg equations ( [ bdg_start ] ) we obtain two independent set of equations @xmath44 and @xmath45 where @xmath46 and @xmath47 . assuming that the system is infinite in the @xmath40 and @xmath41 directions ( @xmath48 ) , the order parameter @xmath49 can be expressed as @xmath50 \nonumber \\ & + & u_\nu ^\downarrow ( z)v_\nu^{\uparrow * } ( z ) f(e^{\downarrow}_\nu ) \ } \label{delta},\end{aligned}\ ] ] where @xmath51 is the electron - phonon coupling and @xmath52 is the fermi - dirac distribution .
the summation in eq .
( [ delta ] ) is carried out only over these states for which the single - electron energy @xmath53 satisfies the condition @xmath54 , where @xmath55 is the debye frequency and @xmath56 is given by @xmath57 where @xmath58 is the nanofilm thickness in the @xmath43-direction .
the system of equations ( [ bdg1da])-([bdg1db ] ) and equation ( [ delta ] ) are solved in a self consistent manner until the convergence is reached . as a result the spatially varying energy gap @xmath49 is obtained . in the further analysis
we often use spatially averaged energy gap defined as @xmath59 since the chemical potential for nanostructures strongly deviates from the bulk value , for each nanofilm thickness we determine the chemical potential by using the formula @xmath60\bigg \}.\end{aligned}\ ] ] in the calculations we adopt the the hard - wall potential profile as the boundary conditions in the @xmath43 direction
. + the set of self - consistent equations can lead to solutions with @xmath61 even for the values of the magnetic field for which the superconducting phase is already not stable - its free energy is greater than the free energy corresponding to the normal metal solution ( @xmath62 ) .
that is why in determining the critical field one should calculate and compare the free energy of the normal and superconducting phase , as it is done here .
in this section we analyze the superconductor to normal metal transition induced by the parallel magnetic field . the thickness range under consideration
is chosen based on the experiments which present measurements for the pb nanofilms with thickness varying from 5 to 30 ml .
the calculations have been carried out for the following values of the parameters : @xmath63 where @xmath64 is the bulk density of the single - electron states at the fermi level , @xmath65 mev , the lattice constant @xmath31 nm and the bulk critical temperature @xmath66 k which corresponds to the energy gap @xmath67 mev . within the parabolic band approximation used in the present model ,
the fermi level in the bulk @xmath68 is treated as a fitting parameter .
@xcite its value is determined based on the experimental results from the photoemission spectroscopy.@xcite we take on @xmath69 ev which corresponds to the @xmath70 nm being about four times the single monolayer thickness @xmath71 . in fig .
[ fig1 ] the in - plane critical field @xmath72 as a function of the nanofilm thickness is presented for different temperatures @xmath73 .
the value of @xmath72 is determined based on calculations of the energy gap as a function of the magnetic field and is defined as the field for which the spatially averaged energy gap @xmath74 drops below @xmath75 . as a function of the nanofilm thickness @xmath58 calculated for the temperatures @xmath76 and
@xmath77 k. results for even - layered nanofilms are marked by triangular red points while for odd - layered nanofilms by square blue points . ]
[ fig1](a ) displays the oscillatory behavior of the paramagnetic critical field with the period of @xmath78 ml . in the thickness range
@xmath79 ml the critical field for the even - layered films is higher than the critical field for the odd - layered films .
this relation reverses for films with the number of monolayers varying from 10 to 22 ml and changes again in the range @xmath80 ml .
the periodicity of such modulation is equal to @xmath81 ml .
the even - odd ( bilayer ) oscillations with the beating effect have been observed in recent measurements of the critical temperature @xcite and the electron - phonon coupling in pb thin nanofilms .
@xcite in these experiments the modulation periodicity varies from 7 ml in ref . to 9 ml in ref . indicating that this quantity depends on the quality of the nanofilm and might vary from one experiment to another .
both the oscillations of @xmath72 in fig .
[ fig1 ] result from the confinement of the electron motion in the direction perpendicular to the film . in the ultra thin nanofilms ,
the confinement of the electron motion leads to the quantization of its energy .
the fermi sphere transforms into the series of parabolic subbands , positions of which on the energy scale decreases with increasing nanofilm thickness . since the cooper - pairing in the phonon - mediated superconductor
is determined by the number of the electron states in the energy window @xmath82 $ ] ( @xmath83 is debye energy ) , the superconducting energy gap increases each time , the subsequent subband passes through the fermi surface .
the presented mechanism , predicted theoretically in 1963 by blatt and thomson @xcite , results in the tooth - like oscillations of the spatially averaged energy gap @xmath74 as a function of the nanofilm thickness .
the tooth - shape of the energy gap oscillations is directly related to the changes of the electron density of states in the vicinity of the fermi level .
it abruptly increases when the subband minima reach the fermi level and then exponentially decreases with increasing nanofilm thickness .
@xcite the zero temperature energy gap @xmath74 as a function of the number of monolayers for pb nanofilms is presented in fig .
[ fig2 ] . as a function of the nanofilm thickness @xmath58 .
results for even - layered nanofilms are marked by triangular red points while for odd - layered nanofilms by square blue points .
green horizontal line denotes the value of the energy gap in the bulk . ]
we see that the shape of @xmath72 oscillations at @xmath84 k [ fig .
[ fig1](a ) ] is directly related to the shape of @xmath74 oscillations in fig .
it is because , in the ultra thin nanofilms with in - plane magnetic field , the pair - breaking mechanism is mainly governed by the paramagnetic ( pauli ) limit which is given by clogston - chandrasekhar @xcite ( cc ) formula , @xmath85 .
therefore , the bilayer oscillations of the in - plane critical field result from the contribution of the subsequent subbands to the superconducting energy gap . as a function of the wave vector @xmath86 for the film thickness ( a ) @xmath87 ml , ( b ) @xmath88 ml , ( c ) @xmath89 ml , ( d ) @xmath90 ml . ] in fig .
[ fig3 ] the quasi - particle energy @xmath91 versus the wave vector @xmath86 is presented for the film thickness @xmath92 ml ( a , b ) and @xmath93 ml ( c , d ) .
the first pair of thicknesses correspond to the range in which @xmath72 is higher for the even - layered films while the second pair is related to the range for which a reversed situation is observed .
we see that for the film thickness @xmath87 ml three lowest subbands participate in the creation of the superconducting state [ fig .
[ fig3](a ) ] . increasing
the thickness by one monolayer causes that the forth subband begins to contribute to the superconducting phase which leads to the enhancement of the energy gap and the critical field depicted in fig .
[ fig1](a ) . in the thickness range
@xmath94 ml the subsequent subbands pass through the fermi level only if the number of monolayers is even .
the period of the passages can be well estimated by the single - electron energy level in the form @xmath95 , where the hard - wall potential is assumed in the @xmath43 direction .
it allows us to determine the period @xmath96 , which for the assumed effective fermi level gives @xmath97 ml . the difference between 2 ml and the real period of the quantum well states which pass through the fermi surface ( @xmath97
ml ) results in the beating effect observed in fig . [ fig1](a ) . in fig .
[ fig3](c , d ) the new subband ( @xmath98 ) appears for the odd - layered film ( @xmath90 ml ) which is in contradiction to the pair @xmath92 ml [ see fig .
[ fig3](a)(b ) ] for which the subsequent subband ( @xmath99 ) appears for the even - layered film ( @xmath88 ml ) .
the interval in which the critical field for the odd - layered films is higher than for the even - layered films ranges from @xmath77 ml to @xmath3 ml .
the source of the beating effect is clearly visible in fig .
[ fig4 ] in which the energy of the quantum well states as a function of the nanofilm thickness is presented . as a function of the film thickness @xmath58 .
results for even - layered nanofilms are marked by triangular red points while for odd - layered nanofilms by square blue points .
the fermi level @xmath100 is set to zero .
the energy closest to the fermi level are joined by the blue(red ) line for the even(odd ) number of monolayers . ] in fig .
[ fig4 ] only the states with the energy below fermi level ( set as zero energy ) are occupied and contribute to the coooper pair condensation .
the closer to the fermi energy they are located the larger is their contribution to the density of states . in fig .
[ fig4 ] the energies closest to the fermi level are joined by the blue ( red ) line for the even ( odd ) number of monolayers .
note that such procedure leads to the `` structure '' , in the sense of the shape , exactly the same as depicted in fig .
[ fig1](a ) .
let us now discuss the value of the paramagnetic critical field obtained from our calculations .
it is well - known that in the ultra thin films with the in - plane magnetic field , the orbital - magnetic interaction is strongly reduced and the upper limit for the critical field is determined by the paramagnetic breakdown of the cooper pairs .
the formula for the paramagnetic critical field in the bulk was derived by clogston and chandrasekhar @xcite and has the form @xmath101 which for pb gives @xmath28 t. this value strongly differs from the critical field obtained in our calculations which vary from @xmath102 t to @xmath103 t. such deviation for the pb nanofilms has been recently reported in ref .
in which the experimental critical field @xmath104 , has been much higher than the paramagnetic limit @xmath105 .
it seems that such discrepancy results from the fact that due to the electron confinement in the nanofilm , the energy gap is enhanced ( fig .
[ fig2 ] ) which leads to increase of the @xmath104 .
nevertheless , the use of the cc formula with the spatially averaged energy gap calculated for each nanofilm thickness @xmath106 still produces the value of the critical field which is lower than the results from numerical solution of the spin - generalized bdg equations , i.e. for @xmath107 ml , @xmath108 t while calculated @xmath109 t. this fact can be understood if we realize that the superconducting energy gap depends on the @xmath43 coordinate and is not uniform as in the expression for clogston - chandrasekhar field .
taking into account the correction which results from the averaging of the energy gap over the @xmath43 coordinate we introduce the thickness - dependent parameter @xmath110 and express the critical field in the form @xmath111 as a function of the film thickness @xmath58 .
the red line is the fitting curve in the form @xmath112 ; @xmath113 .
the inset : the spatially varying energy gap @xmath49 calculated for the nanofilm thickness @xmath107 ml and @xmath114 ml . ] fig .
[ fig5 ] shows that the parameter @xmath115 is the decreasing function of the thickness @xmath58 and for sufficiently large thickness reaches the value @xmath116 .
the asymptotic behavior of @xmath117 for @xmath118 results from the fact that for the large thickness the spatially varying energy gap diverges to the uniform one - compare @xmath49 for the thicknesses @xmath107 and @xmath119 ml presented in the inset of fig .
the dependence @xmath110 can be well fitted using the formula @xmath112 ( the thickness @xmath58 denotes the number of monolayers - see fig .
[ fig5 ] ) which clearly indicates that the parameter @xmath115 diverges to the bulk value @xmath120 in accordance to the cc theory .
it should be noted that the smallest difference between the quasiparticle energy in the superconducting state and the fermi energy is the same for each band created due to the confinement of electrons in the direction perpendicular to the sample ( c.f .
[ fig3 ] ) . using this energy gap , instead of the averaged energy gap @xmath74 in the cc formula ,
one obtains the same values of critical fields as the ones resulting from the bdg equations . in the present subsection
we discuss in detail the effect of temperature on the superconductor to normal metal transition induced by the in - plane magnetic field .
the critical field @xmath121 presented in fig .
[ fig1 ] for different temperatures @xmath73 shows that its value gradually decreases with increasing temperature . since the critical temperature oscillates as a function of the nanofilm thickness ( similarly to the energy gap depicted in fig .
[ fig2 ] ) , at the temperature @xmath122 k , the nanofilms with the thickness @xmath92 ml are superconducting while the film with the thickness therebetween transits to metallic state . in fig .
[ fig6 ] we present the spatially averaged energy gap as a function of the magnetic field and temperature .
the results in fig .
[ fig6 ] have been calculated for the nanofilm thicknesses @xmath123 and @xmath124 ml .
to make it more transparent , the value of the energy gap in each of the figures is normalized with respect to its maximum . as a function of the magnetic field @xmath125 and temperature @xmath73 for the nanofilm thickness ( a ) @xmath87 ml , ( b ) @xmath88 ml , ( c ) @xmath89 ml and ( d ) @xmath90 ml .
the value of energy gap in each figure is normalized with respect to its maximum . ] in fig .
[ fig6 ] we can see that the range of the magnetic field and temperature for which the sample is superconducting is larger for the thickness for which the enhancement of the energy gap ( see fig .
[ fig2 ] ) is found .
nevertheless the thickness - dependence is not visible in the normalized critical magnetic field @xmath126 as a function of @xmath127 , where @xmath128 and @xmath129 is the critical field at @xmath84 and the critical temperature for @xmath130 , respectively . in other words
the phase diagram @xmath131 presented in fig .
[ fig7 ] looks the same for each film thickness which is understandable if we recall our restriction to the pauli limit with no orbital effect . as a function of @xmath127 obtained from the solution of bdg equations . here
the dashed green line corresponds to @xmath132 .
the first to second order transition , @xmath133 , is marked by arrow . ] the initial behavior of @xmath134 at the temperature close to @xmath1 ( @xmath135 ) can be well estimated by the formula @xmath136 .
note that , in contrast to the orbital limit with no pauli effect , the slope of @xmath134 at critical temperature @xmath135 is infinite and in the vicinity of @xmath135 is expressed by @xmath137 . in this range
the transition at @xmath72 is of second order . in the pauli
limit the second order transition is suppressed with decreasing temperature and below @xmath138 it becomes the first order .
@xcite the first to second order transition is clearly presented in fig .
[ fig8 ] in which the energy gap versus magnetic field is presented for different temperatures . as a function of magnetic field @xmath125 for different temperatures .
the first to second order transition emergences at @xmath139 .
calculations performed for the film thickness @xmath89 ml . ]
note that the critical field @xmath72 at @xmath84 is not equal to the paramagnetic limit @xmath105 but is further enhanced .
a detailed discussion of the deviation of the critical field from @xmath105 is presented in sec .
[ sec : results].a .
the microscopic model based on the spin - generalized bdg equations allows to analyze superconductor to normal metal transition in the nanofilms when its thickness is reduced to few nanometers .
however , in the nanoscale regime , the confinement affects not only the electronic spectrum , but also the phononic degrees of freedom . the phonon dispersion in thin films strongly deviates from that in the bulk .
@xcite the quantization of the phononic spectra in the nanofilms and its influence on @xmath1 and energy gap oscillations have been considered in ref . .
however , the effect of the confinement on the electron - phonon coupling strength has not been included .
this lack has been recently supplemented by saniz et al . in ref . .
in this paper @xcite the authors have investigated the effect of confinement on the strength of the electron - phonon coupling as well as the electronic spectrum and its influence on the oscillations of the critical temperature .
the formula for the phonon - mediated attractive electron - electron interaction has been derived with the use of the green function approach beyond the contact potential approximation . it has been found that the increase of the critical temperature observed in superconducting nanofilms is due to the increase of the number of phonon modes what results in the enhancement of the electron - phonon coupling . in contradiction to previous models @xcite , this study predicts the suppression of the critical temperature with increasing density of states at the fermi level .
such conclusion seems not to be confirmed by recent experiments in which the direct correlation between the @xmath1 oscillations and the energy distribution of the quantum well states in nanofilms is observed .
@xcite in order to determine the realistic value of the electron - phonon coupling which occurs in experiments , there is one more aspect which should be taken into consideration .
note that the results presented in this paper regard the so - called freestanding nanofilms .
in fact , pb films are grown on si(111 ) substrate and , for the purpose of the transport measurements , are covered by the layer of au .
both these layers strongly affect the electron - phonon coupling in the nanofilm due to the interface effect .
it has been recently observed @xcite that the electron - phonon coupling for pb nanofilms on si substrate is lower than in the bulk and diverges to the bulk value with increasing film thickness .
the results presented in ref . indicate that the interface effect has a strong impact on the electron - phonon coupling and should be taken into account in the presented model in order to reproduce the experimental data .
the simple approximation which allows to take this effect into account has been recently proposed by chen et al . in ref . .
it has been assumed @xcite that in the first approximation the spatially - dependent electron - phonon coupling can be well estimated by the formula @xmath140 where @xmath141 is the interface thickness with the electron - phonon coupling @xmath142 and @xmath143 is the electron - phonon coupling in the bulk .
averaging over the @xmath43 coordinate and including the oscillatory behavior of @xmath51 coefficient reported in experiments @xcite leads to the formula ( for details see ref . )
@xmath144 where @xmath145 is a periodic function , @xmath146 is the number of monolayers , @xmath147 is the fermi wavelength and @xmath71 is the lattice constant . since the period of @xmath1 oscillations in pb nanofilms is measured to be @xmath148 ml , the authors of ref .
have reduced the function @xmath145 to only two parameters : @xmath149 for an odd number of monolayers and @xmath150 for an even number of monolayers . by appropriate choice of these parameters and by setting the fermi level to the value corresponding to the period of the quantum - size oscillations ( @xmath78 ml )
one can well reproduce the experimental data of @xmath151 coming form ref . .
the values of the parameters used for each of the mentioned references is presented in table [ tab1 ] .
.the value of parameters used in ref . to reproduce the oscillations of @xmath1 observed in experiments -
see reference column . [
cols="^,^,^,^",options="header " , ] in the present subsection we use these parameters to predict the zero temperature in - plane critical field as a function of the nanofilm thickness .
we believe that such procedure allows to determine the realistic value of the critical field comparable with recent experimental measurements .
@xcite the thickness - dependent oscillations of the zero temperature critical field are presented in fig .
although the periodicity of the @xmath72 oscillations for each of the considered references is the same and equal 2 ml , the characteristics of these oscillations are different and vary form one experiment to another .
this difference regards the overall trend of @xmath72 which is an increasing function of the thickness in case of refs .
but decreasing in the case of ref . .
note that the values of the critical field in fig .
[ fig9 ] are suppressed as compared to the critical field calculated with the electron - phonon coupling for the bulk [ compare with fig .
[ fig1](a ) ] .
the superconductor to normal metal phase transition driven by the magnetic field for pb nanofilms is investigated with the use of the bdg equations .
only the pauli pair - breaking mechanism is included as it is assumed that the external magnetic field is parallel to the surface of the nanofilm .
it is shown that the even - odd oscillations of the critical magnetic field appear as the thickness of the sample is increased .
the shape of the @xmath72 oscillations is directly related to the shape of the thickness dependence of the superconducting gap . the beating effect visible in the oscillations
is discussed in the context of the energy of the quantum well states which appears due to the confinement of the electron motion in the direction perpendicular to the sample . as it has been shown
the period in the nanofilm thickness between two neighbouring peaks in the critical field is equal to 2.14 ml . as the number of the monolayers has to be an integer the beating effect appears in the oscillatory behavior of @xmath72 and @xmath74 .
we also show that the zero - temperature critical field in the nanofilms is higher than the clogston - chandrasekhar paramagnetic limit and it diverges to the cc value for sufficiently thick films .
this fact is explained in term of the spatially varying energy gap induced by the confinement .
the phase diagrams in the @xmath152 plane are presented for different values of @xmath58 . according to the obtained results the thickness - dependence is not visible in the phase diagrams when both the field and the temperature are normalized @xmath153 and @xmath127 .
moreover , the interface effect on the electron - phonon coupling is included by using a simple approximation proposed in ref . .
this leads to a spatially dependent @xmath51 factor and its oscillatory behavior with increasing nanofilm thickness .
as it is shown this effect leads to a suppression of the critical field values in comparison to the ones corresponding to the bulk .
we believe that such approach allows for comparison of the calculated @xmath72 oscillations with the experimental results .
discussions with jzef spaek are gratefully acknowledged .
this work was financed from the budget for polish science in the years 2013 - 2014 .
project number : ip2012 048572 .
m. z. acknowledges the financial support from the foundation for polish science ( fnp ) within project team .
32ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | the quantum size effect on the in - plane paramagnetic critical field in pb nanofilms is investigated with the use of the spin - generalized bogolubov - de gennes equations .
it is shown that the critical field oscillates as a function of the nanofilm thickness with the period of @xmath0 ml ( even - odd oscillations ) modulated by the beating effect .
this phenomena is studied in terms of the quantization of the electron energy spectra caused by the confinement of the electron motion in the direction perpendicular to the sample .
the calculated values of critical fields for different nanofilm thicknesses are analyzed in the context of clogston - chandrasekhar limit .
the influence of the thermal effect on the magnetic field induced superconductor to normal metal transition is also discussed .
furthermore , the thickness - dependence of the electron - phonon coupling and its influence on the value of the critical magnetic field are studied . |
although the structural sequence of the high - pressure phase transitions in si and ge has been considered to be well known since two decades , a new high - pressure phase was found experimentally between the body - centered tetragonal structure with two basis atoms ( which will be referred to the following as @xmath1-tin or bct phase ) and the simple - hexagonal ( sh ) structure in si@xcite and ge.@xcite after its space group this intermediate phase was called _ imma _ phase and corresponds to a body - centered orthorhombic bravais lattice with two atoms in the unit cell ( which will be called in the following bco ) .
thus , the experimental structural sequence is found to be cd @xmath0 @xmath1-tin @xmath0 _ imma _ @xmath0 sh @xmath2 , where cd indicates the cubic - diamond structure .
a bco structure was first theoretically proposed for si by needs and martin@xcite about twenty years ago , whereas lewis and cohen@xcite predicted exactly the _ imma _ phase for ge a few years before the experimental confirmation .
the first theoretical investigations for si and ge indicated for both , the @xmath1-tin@xmath0_imma _ and the _ imma_@xmath0sh case , a continuous second - order phase transition.@xcite at variance , the experiments show first - order phase transitions with a discontinuity of the volume in both substances and a hysteresis effect in si.@xcite for ge the first - order transition was detected only for the @xmath1-tin@xmath0_imma _
transition,@xcite because the _ imma_@xmath0sh transition was expected to be beyond the experimentally accessible pressure range .
the transition pressures in ge are expected to be higher than the corresponding ones in si , because of the strong repulsive character of the atomic potential in ge due to the presence of @xmath3 electrons in the core.@xcite from the theoretical point of view , recent total - energy calculations show two first - order phase transitions @xmath1-tin@xmath0_imma _ and _ imma_@xmath0sh both for si@xcite and ge.@xcite a second - order phase transition was found for the @xmath1-tin@xmath0_imma _ case from elastic instabilities.@xcite first information on the order of the phase transition can be given by group - theoretical arguments:@xcite if the order of the point group of one phase is one half of the order of the point group of the other phase , then the phase transition can be of second order .
if the point group of one phase is one third of the order of the point group of the other phase , then the phase transition must be of first order .
the order of the point group @xmath4 of the cd structure ( space group @xmath5 , @xmath6 ) is 48 . since the order of the point group @xmath7 of the @xmath1-tin phase ( space group @xmath8 , @xmath9 ) is 16
, the cd@xmath0 @xmath1-tin transition has to be of first order .
the order of the point group @xmath10 of the _ imma _ phase ( space group @xmath11 , @xmath12 ) is 8 , thus the @xmath1-tin@xmath0_imma _ transition can be a second - order one .
furthermore , the order of the point group @xmath13 of the sh structure ( space group @xmath14 , @xmath15 ) is 24 , the sh@xmath0_imma _ and also the _ imma_@xmath0sh transitions are of first order . following the group - theoretical analysis none of the available total - energy calculations
have described correctly the order for the phase transitions of the whole sequence , neither for si nor for ge .
this paper is organised as follows : in section [ method ] we give a short overview of the methods on which our calculations are based . in section
[ parameters ] we report the results for the equilibrium parameters of these phases and compare them with the experimental volume dependence .
transition pressure ranges are presented in section [ pressure ] .
the order of the phase transitions is analysed in section [ order ] , and the enthalpy barriers between the phases are determined in section [ barriers ] .
all results are summed up and discussed in section [ discussion ] .
the conclusions are summarised in section [ conclusion ] .
our calculations have been performed using the plane - wave pseudopotential approach to the density - functional theory ( dft)@xcite implemented within the vienna _ ab - initio _ simulation package ( vasp).@xcite for the exchange - correlation potential we have used the generalised - gradient approximation ( gga ) by perdew and wang@xcite for si and ge , and , because of poorer results , the local - density approximation ( lda)@xcite for si only .
the interaction with the ion cores has been described by ultrasoft pseudopotentials.@xcite the wave functions have been expanded in terms of plane waves up to a kinetic - energy cutoff of 270 ev ( 410 ev ) for si ( ge ) .
this choice provided an error smaller than 0.5 kbar ( 0.2 kbar ) for si ( ge ) to the pressure according to the pulay stress.@xcite the pulay stress arises from using an incomplete plane - wave basis set and causes errors in the stress tensor and , correspondingly , in the pressure .
convergence also required a 18@xmath1618@xmath1618 ( 24@xmath1624@xmath1624 ) mesh of monkhorst - pack points@xcite which yielded 864 ( 1962 ) * k*-points in the irreducible wedge of the brillouin zone for si ( ge ) .
all high - pressure phases are metallic .
thus , we have used a methfessel - paxton smearing@xcite with a width of 0.2 ev . in order to determine the total energy for a given volume or structure we have minimised the energy with a conjugate - gradient algorithm.@xcite the equilibrium properties
have been obtained by fitting the values of the energy as a function of the volume alternatively to the vinet@xcite and to the murnaghan@xcite equation of state .
in this section we present our results for the volume dependence of the structural parameters and compare them with the experimental ones . the structural parameters are more useful indicators of phase transitions than the often - used tangent construction to the equation - of - state curves , the difficulties of which will be illuminated .
we have focussed our attention to the structural sequence @xmath1-tin@xmath0_imma_@xmath0sh .
the calculations for the cd phase have been performed just for the sake of comparison .
the most general structure which is compatible with all four phases is the body - centered orthorhombic cell with a basis of two atoms at ( 0,0,0 ) and ( 0,@xmath17,@xmath18 ) ( bco structure ) .
the other structures are special cases : for the bct structure ( @xmath1-tin ) one has the lattice constants @xmath19 and the internal parameter @xmath20 ; for the sh structure one has @xmath21 and @xmath22 ; for the cd phase one has @xmath23 , @xmath24 , and @xmath20 .
the pure _ imma _ phase is realised as a bco structure with @xmath25 and @xmath26 . the minimization of the total energy has been performed using the following procedure : for all phases we have started from the bco structure and minimised the energy as a function of the structural parameters with the constraints appropriate to the various phases .
in this way we have avoided an energy offset which arises from using different cells .
for example , in the case of the cd structure we have imposed , according to the previous considerations , @xmath23 , @xmath24 , and @xmath20 , and then we have minimised the total energy as a function of @xmath27 . in the special case of the sh structure
, we could keep only the value of @xmath28 fixed at 0.5 , and we have relaxed the value of @xmath29 ( and also @xmath30 ) by minimization of the energy ; the structure which is obtained after relaxation will be indicated in the following as sh . indeed , converged calculations should be yield @xmath31 ( sh @xmath32 sh ) after relaxation . for the _ imma _ phase , all lattice and internal parameters of the bco structure were allowed to relax
. the relaxed parameters as a function of the volume are displayed in figs .
[ relaxsi ] and [ relaxge ] for si and ge , respectively . the results from calculations performed withing the gga and the lda are quite similar for si .
since it turned out , that the pressures calculated by gga are more accurate in si , only gga calculations have been performed for ge . as can be seen in fig .
[ relaxsi ] , the relaxation of @xmath29 for the sh phase of si does not match the ideal value . at variance , the calculation for ge does reproduce this value .
this indicates that for a full convergence a higher kinetic - energy cutoff and/or a larger grid of special points should have been used for si .
however , the condition @xmath31 is almost fulfilled in the region of stability of the sh phase .
furthermore , the energy obtained from a calculation in the same cell with @xmath21 fixed and @xmath30 relaxed differs from the value obtained using the previous procedure by less than 0.7 mev .
for a more detailed discussion see ref . .
the agreement between our data and those experimental ones which are not included in the figure is good : the reported value of @xmath33 for the @xmath1-tin phase of si@xcite is well within the range of our values .
the experimental values of @xmath34 and @xmath35 at the volume @xmath36 @xmath37 of ref . agree with our data .
furthermore , the measured value of 24.22 @xmath37 for the volume of the sh phase of si@xcite is also in the stability range we have found for this phase . for ge in the bct structure , a @xmath30 ratio of 0.5480.554 was detected experimentally@xcite corresponding to a volume region around 33 @xmath37 ; in this region , our @xmath30 ratio is 0.5460.549 . for the sh phase of ge , a @xmath38 ratio of 0.93 was measured@xcite which also agrees with our data .
figures [ relaxsi ] and [ relaxge ] show that the relaxation of the general bco structure leads to the bct structure for large volumes and to the sh structure for small ones . in the intermediate region , extending from about 26.5 to 28.0 @xmath37 for si and 23.5 from 25.5 @xmath37 for ge , we localise the stability range of the pure _ imma _ phase .
whereas the transition between the _ imma _ and the sh phase is characterised by a discontinuity in most of the structural parameters , the _ imma _ and the @xmath1-tin phase are quite difficult to distinguish by an inspection of figs .
[ relaxsi ] and [ relaxge ] .
a clearer picture is obtained from the electronic structure and the fermi energy ( see appendix [ app ] for details ) . from this procedure , the stability range of the _ imma _ phase is from 26.6 to 28.3 @xmath37 for si ( gga ) , from 26.6 to 28.0 @xmath37 for si ( lda ) , and from 23.7 to 25.4 @xmath37 for ge .
the standard procedure for calculating the volume change and the pressure at the transition is the so - called common - tangent construction .
it relies on the fact that , in the case of a first - order phase transition , two structures with the same pressure @xmath39 and enthalpy @xmath40 but different total energy @xmath41 coexist .
the volume derivative of the total energy is the negative pressure .
therefore , the slope of the common tangent to the two energy curves gives the pressure for the transition between the corresponding structures .
due to the fact that calculations are performed at a finite number of volumes , a fit of an analytical expression to a continuous energy curve is needed in order to obtain a continuous derivative . in fig .
[ ev ] we show the energy of the relaxed structures for si as a function of volume .
the volume region where the pure _ imma _ phase is stable does not include any minimum of the total energy @xmath42 .
this makes a fit to any equation of state for the _ imma _ phase slightly unstable , which results in an unreliable determination of the transition pressures and of the volume change through the common - tangent construction in this region .
the results from a fit of the vinet@xcite and murnaghan@xcite equation of state are shown in table [ eos ] for ge as well as a comparison of the lda and gga calculations for si .
.equilibrium parameter determined by the murnaghan ( mur ) and vinet ( vin ) equation of state for the cd , @xmath1-tin , _ imma _ , and sh phase of si and ge .
the equilibrium volume @xmath43 is in units of @xmath37 and the bulk modulus @xmath44 in kbar . [
cols="<,>,^,^,^,^,^,^ " , ] for si the value of the pressure for the cd@xmath0@xmath1-tin transition calculated with gga is within the range of the experimental values whereas the lda calculation underestimates it .
we could not reproduce the experimentally observed hysteresis@xcite for the @xmath1-tin@xmath0_imma _ transition in si .
for a first - order phase transitions , such as _ imma_@xmath0sh , a hysteresis could occur due to the behaviour of the hydrostatic line as in fig .
[ gebcoh ] .
since our results support a second - order phase transition for @xmath1-tin@xmath0_imma _ rather than a first - order one , no hysteresis should be found in this case .
the transition pressures mentioned in ref . are obtained either from the first appearance of a new phase or from the complete transition of the sample from one phase to the other .
thus we estimate the experimental range in which the @xmath1-tin@xmath0_imma _ and _ imma_@xmath0sh transitions occur in analogy to the procedure explained in section [ pressure ] .
unfortunately , the calculated transition pressures for @xmath1-tin@xmath0_imma _ lie below the ones for the first cd@xmath0@xmath1-tin phase transition and underestimate the experimental ones .
the next phase transition to be considered is _ imma_@xmath0sh . here
a hysteresis was found within our calculations as well as in the experiment .
the results for the transition pressures within lda underestimate again the experimental ones , whereas the ones within gga overestimate them .
not included in table [ summary ] are the experimental pressures for mixed ( 50:50 ) _ imma_-sh samples , which are 154 kbar for pressure increase and 159 kbar for pressure decrease@xcite .
this is the only case in which the transition to the ( mixed ) _ imma _ phase occur at a lower pressure for an increasing than for a decreasing external pressure .
all other experimental points show the same kind of hysteresis which is reproduced in our results .
the difference of the pressures , where the _ imma _ phase was measured with pressure increase and decrease is 5 kbar for the mixed samples and 1112 kbar for the pure ones . from the experimental transition
pressures the hysteresis is found to be between 3 and 14 kbar .
within our results the largest possible hysteresis is found to be around 29 ( 24 ) kbar and the smallest around 112 ( 27 ) kbar for gga ( lda ) , which is in agreement with the experimental values .
the transition pressures determined by the crossing of the enthalpy curves are found to be within the hysteresis range calculated here . for ge
the transition pressure for the cd@xmath0@xmath1-tin transition is slightly smaller than the experimental one but the results for the @xmath1-tin@xmath0_imma _ transition match much better . in the experiment@xcite
just the @xmath1-tin@xmath0_imma _ transition was examined , because the transition pressure for the _ imma_@xmath0sh transition was higher than the maximum accessible pressure of 810 kbar . in comparison to the experimental results of si
it is assumed that the _ imma _ phase for ge is stable from 750 to 850 kbar , which is compatible with the observation of a sh phase at 900 kbar.@xcite from this point of view , the transition pressure calculated here is slightly too high .
there are many reasons for the discrepancy between our results and the experimental ones . on one hand ,
the experimental conditions like pressure profiles or measurement times are not perfectly controlled , whereas nonhydrostaic conditions affect the transition pressures ; on the other hand , temperatue effects , which are neglected in our calculation , could be of some relevance .
in fact , for si the theoretical transition pressure decreases with increasing temperature.@xcite
we have presented a first - principles investigation of the pressure - induced phase transitions cd @xmath0 @xmath1-tin @xmath0 _ imma _ @xmath0 sh in si and ge . numerical accuracy has required these calculations to be performed within the same unit cell and the pressure to be calculated from the stress tensor instead of using any equation of state .
we have determined the phase transitions by investigating the relaxed parameters , the enthalpy vs. pressure curves , and the hydrostatic condition . from the pressure dependence of the relaxed structure parameters the hysteresis effect does not seem to have been verified in a theoretical investigation before .
an examination of the hydrostatic condition provides a thermodynamical explanation for the existence of this hysteresis .
further on , the orders of the phase transitions have been investigated with the helps of the curves of the enthalpy vs. pressure and of the derivative of the enthalpy vs. pressure .
for si the transitions _ imma_@xmath0sh and @xmath1-tin@xmath0_imma _ are found to be of first and of second order , respectively . for ge
the results support a second - order @xmath1-tin@xmath0_imma _ transition more than a first - order one , but in this case the order can not be definitely determined .
a calculation of the enthalpy barriers between the phases show a vanishing barrier between the @xmath1-tin and the _ imma _ phase for si , which also supports the second - order @xmath1-tin@xmath0_imma _ transition , whereas no barrier could be found for all @xmath1-tin@xmath0_imma_@xmath0sh transitions .
finally , our alternative results are fully consistent with each other and show an acceptable agreement with the available experimental data .
support by the heinrich bll stiftung , germany , is gratefully acknowledged .
we examined the volume dependence of the fermi energy of bct , bco , and sh . in fig .
[ fermi ] we show the fermi energy reduced by the bco values for ge , @xmath45 in the range of stability of the @xmath1-tin phase the fermi energy of bco is identical with the one of bct .
similarly , the fermi energy of bco and sh are identical in the range of stability of the sh phase .
therefore , the regime where the fermi energy of bco is neither equal to the one of bct nor to the one of sh is assumed here to be the range of stability of the pure _ imma _ phase .
the energy values @xmath47 are reduced in such a way that the fermi energy at any volume is set equal to zero , @xmath48 in most of the volume range presented in fig .
[ eigen ] the band structure presents a linear dependence on volume .
however , the energy gradients differ for each structure . within the transition region between the _ imma _ and the sh phase one observes a small discontinuity ( here apparent at @xmath49@xmath37 ) which is attributed to the first - order phase transition . when starting from the @xmath1-tin phase , a deviation of the bco band energies from the bct energies can be observed , which increases continuously .
thereby , the degeneracy of the bct bands is lifted , which indicates the appearance of the less symmetrical _ imma _ phase .
the behavior of the band structure for si is very similar to the one of ge . | we have investigated the structural sequence of the high - pressure phases of silicon and germanium .
we have focussed on the cd @xmath0 @xmath1-tin @xmath0 _ imma _ @xmath0 sh phase transitions .
we have used the plane - wave pseudopotential approach to the density - functional theory implemented within the vienna _ ab - initio _ simulation package ( vasp ) .
we have determined the equilibrium properties of each structure and the values of the critical parameters including a hysteresis effect at the phase transitions .
the order of the phase transitions has been obtained alternatively from the pressure dependence of the enthalpy and of the internal structure parameters .
the commonly used tangent construction is shown to be very unreliable .
our calculations identify a first - order phase transition from the cd to the @xmath1-tin and from the _ imma _ to the sh phase , and they indicate the possibility of a second - order phase - transition from the @xmath1-tin to the _ imma _ phase . finally , we have derived the enthalpy barriers between the phases . |
one of the problems originally presented in _ what is mathematics ? _
book by courant and robbins @xcite is stated as follows : _ suppose a train travels from station @xmath0 to station @xmath1 along a straight section of track .
the journey need not be of uniform speed or acceleration .
the train may act in any manner , speeding up , slowing down , coming to a halt , or even backing up for a while , before reaching @xmath1 .
but the exact motion of the train is supposed to be known in advance ; that is , the function @xmath2 is given , where @xmath3 is the distance of the train from station @xmath0 , and @xmath4 is the time , measured from the instant of departure .
on the floor of one of the cars a rod is pivoted so that it may move without friction either forward or backward until it touches the floor .
if it does touch the floor , we assume that it remains on the floor henceforth ; this will be the case if the rod does not bounce .
is it possible to place the rod in such a position that , if it is released at the instant when the train starts and allowed to move solely under the influence of gravity and the motion of the train , it will not fall to the floor during the entire journey from @xmath0 to @xmath1 ? _ as an exercise , the authors suggest to prove a more general result where a spherical inverted pendulum is considered instead of a planar pendulum .
we consider slightly different problem introducing a viscous friction force acting on the mass point of the pendulum and prove that for an arbitrarily small friction coefficient a periodic solution always exists .
moreover , we show that the considered periodic solution never approaches horizontal plane . in the first section ,
we present the system of governing dynamical equations for the system and prove some properties which we are going to use in the second section where we apply some recent results by srzednicki , wjcik and zgliczynski @xcite to our system .
let @xmath5 be an orthogonal moving reference frame such that @xmath6 coincide with the pivot point of the pendulum , @xmath7 and @xmath8 axis are in a horizontal plane and always remain parallel to themselves at the initial moment of time ; @xmath9 is vertical and oriented in an opposite way to the gravitational force . by @xmath10 we denote the radius vector of the mass point in @xmath5 .
let @xmath11 , @xmath12 , and @xmath13 be its components .
since the pivot point moves periodically in the horizontal plane , then for the radius vector @xmath14 of the mass point in some fixed reference frame with its axis parallel to the axis of @xmath5 , we have @xmath15 where @xmath16 ; we assume @xmath17 to be @xmath18-periodic smooth functions .
obviously , we have similar relations for the velocities and accelerations @xmath19 the mass point moves under the action of the gravitational force , the viscous friction force , and the force of constraint @xmath20 here @xmath21 is the constraint force which is parallel to the radius vector @xmath10 .
@xmath22 forces of constraint are the forces which allows one to consider a system with constraints as constraint - free with additional unknown _ apriori _ forces .
more on constraint forces and their use in mechanics one can find in @xcite .
we assume that the friction force is determined by the following model @xmath23 we can now rewrite ( [ eq1 ] ) as follows @xmath24 for the above equation ( [ eq1mod ] ) , the solution function @xmath25 is determined by the initial conditions @xmath26 and @xmath27 . since functions @xmath28 and @xmath29
are @xmath18-periodic , then the extended phase space of our system can be considered as @xmath30 .
let @xmath31 be a function defined by the following equation @xmath32 now consider the submanifold @xmath33 of the extended phase space and show that if @xmath34 is large , then along the solutions starting at @xmath33 the function @xmath35 is locally decreasing .
more specifically , [ lem1 ] there exists @xmath36 such that @xmath37 from the definition of @xmath35 and ( [ eq1 ] ) we easily have @xmath38 therefore , taking into account ( [ eq2 ] ) and that @xmath39 , we obtain @xmath40 since @xmath41 is bounded , then for @xmath34 sufficiently large we have @xmath42 .
now we prove that if the pendulum , as well as its velocity vector , is in the horizontal plane , then at least locally the mass point is falling down , i.e. the following lemma is true .
[ lem2 ] if the solution @xmath10 of ( [ eq1 ] ) at the time @xmath4 satisfies @xmath43 and @xmath44 , then @xmath45 . for the pendulum being in the horizontal position and moving horizontally as well
, we have the following : 1 .
the friction force is always parallel to the radius vector @xmath46 , therefore , at the given conditions , the friction force is directed horizontally ; 2 .
the constraint force is in the horizontal plane if @xmath47 ; 3 .
@xmath5 is moving along the horizontal plane , therefore , @xmath48 = 0 .
taking into account the above , from ( [ eq1 ] ) we obtain @xmath49
in this section we are going to apply some recent developments in fixed point theory by srzednicki , wjcik and zgliczynski @xcite to our system and prove the existence of a falling - free periodic solution .
first , following @xcite , we introduce some definitions which we slightly modify for our use .
* there exists a compact subset @xmath75 of @xmath58 called the essential exit set such that @xmath76 * there exists a homeomorphism @xmath77\times w_a \to w$ ] such that @xmath78 and @xmath79\times w_a^{--})=w^{--}.\ ] ] the monodromy map @xmath81 is a homeomorphism .
moreover , it can be proved that a different choice of @xmath83 satisfying [ cond-2 ] leads to the monodromy map homotopic to @xmath81 .
it follows that the isomorphism in homologies @xmath84 is an invariant of @xmath65 .
[ th1 ] @xcite let w be a periodic segment over @xmath74 $ ] .
then the set @xmath85 \}\ ] ] is open in @xmath86 and the set of fixed points of the restriction @xmath87 is compact . moreover ,
if @xmath65 and @xmath75 are anrs then @xmath88 where by @xmath89 and @xmath90 we denote the lefschetz number of @xmath81 and @xmath91 respectively . in particular , if @xmath92 then @xmath93 has a fixed point in @xmath86 .
let @xmath99 be a manifold with a boundary defined by the inequalities @xmath100 , @xmath101 , where @xmath102 is obtained from lemma [ lem1 ] .
one can easily prove that @xmath65 is diffeomorphic to @xmath103 , where @xmath104 is a 2-dimensional disk ( with boundary ) .
let us now prove that @xmath65 is a periodic segment over @xmath105 $ ] for ( [ eq1 ] ) .
indeed , from lemma [ lem1 ] we see that the essential exit set is entirely on the boundary @xmath47 .
moreover , from lemma [ lem2 ] we obtain that the points which satisfies @xmath47 and @xmath106 do not belong to the essential exit set , at the same time , if for some point we have @xmath47 , @xmath107 then this point is in the essential exit set
. therefore , the essential exit set is compact .
homeomorphism @xmath83 in our case is an identity map , i.e. @xmath105\times w_0 = w$ ] .
finally , one can easily see that @xmath108 is homotopic to @xmath109 and @xmath110 is homotopic to @xmath111 .
therefore , taking into account that @xmath112 , we get @xmath113 and theorem [ th1 ] can be applied .
r. courant , h. robbins , _ what is mathematics ? : an elementary approach to ideas and methods _ , oxford university press ( 1996 ) .
r. srzednicki , k. wjcik , and p. zgliczynski , _ fixed point results based on waewski method _ in _ handbook of topological fixed point theory _ , ed : r. brown , m. furi , l. grniewicz , b. jiang ( 2005 ) , 903941 .
h. goldstein , c. poole , j. safko , _ classical mechanics _ ( 2002 ) . | for the system of an inverted spherical pendulum with friction and a periodically moving pivot point we prove the existence of at least one periodic solution with the additional property of being falling - free .
the last means that the pendulum never becomes horizontal along the considered periodic solution .
presented proof is an application of some recent results in the fixed point theory . |
the formation of new blood vessels ( angiogenesis ) is the essential mechanism for tumour progression and metastasis .
a simplified model of the tumor angiogenesis can be described by a keller - segel equation @xcite . in this paper
we consider the equation on the two dimensional cylindrical domain @xmath4 $ ] with a front boundary condition in @xmath5 and the periodic condition in @xmath6 , both specified later , @xmath7 with @xmath8 , where @xmath9 denote the density of endothelial cells , @xmath10 stands for the concentration of the chemical substance or the protein known as the vascular endothelial growth factor ( vegf ) and @xmath11 is a decreasing chemosensitivity function , reflecting that the chemosensitivity is lower for higher concentration of the chemical .
the system includes the both zero chemical diffusion @xmath12 and non - zero chemical diffusion @xmath13 cases .
+ endothelial cells forming the linings of the blood vessels are responsible for extending and remodeling the network of blood vessels , tissue growth and repair .
tumors or cancerous cells are also dependent on bloods supply by newly generated capillaries formed toward them , which process is called the endothelial angiogenesis . in the modeling the endothelial angiogenesis , the biological implication is that the endothelial cells behaves as a invasive species , responding to signals produced by the tissue . accordingly
the system is given the front condition at left - right ends such that @xmath14 we choose the @xmath5-axis by the propagating direction .
a _ planar _ traveling wave solution of is a traveling wave solution independent of the transversal direction @xmath6 such that @xmath15 with a wave speed @xmath16 . from now on ,
we consider only planar traveling waves @xmath17 satisfying the above boundary conditions and , and moreover we assume @xmath18 ( we denote @xmath19 by @xmath20 in short . )
+ in this paper we study the stability of a planar traveling wave solution @xmath17 of for @xmath21 case assuming @xmath22 , @xmath23 , @xmath24 .
\end{aligned } \end{aligned}\ ] ] main theorems let us introduce the weight function : @xmath25 and define the sobolev spaces @xmath26 and @xmath27 by their norms : @xmath28 } | { \partial}_x^i { \partial}_y^j \varphi ( x , y ) |^2 dxdy , \quad\\ & \| \varphi\|_{h^k_w}^2 = \sum _ { i+ j \le k}\int_{{\mathbb r}\times [ 0 , { \lambda } ] } | { \partial}_x^i { \partial}_y^j \varphi ( x , y ) |^2 w(x ) dxdy , \quad k = 0 , 1 , 2 \dots .
\end{aligned}\ ] ] we will assume the periodic boundary condition in @xmath6-direction for initial perturbation and use the sobolev spaces @xmath29 and @xmath30 : @xmath31 @xmath32 where @xmath2 is the circle of perimeter @xmath3 and @xmath33 is the @xmath34th fourier coefficient of @xmath35 .
+ in other words , @xmath36 is a function on @xmath37 such that it is @xmath3-periodic in @xmath6 and its ( weighted in @xmath38-direction ) @xmath26 norm on @xmath39 $ ] is finite .
the cylindrical domain with this periodic boundary condition fits well to the picture that blood vessels are lined with a single layer of endothelial cells .
what it follows we use the equivalence of @xmath40 and @xmath41 norms to @xmath42 and @xmath43 norms for functions periodic in @xmath6 .
+ we perturb the system around a planar traveling wave solution @xmath17 with the front conditions , and in a way that @xmath44 where the traveling wave speed @xmath45 . what it follows , we will use the moving frame @xmath46 we consider the above perturbation @xmath47 in the following function class . 1 .
there exists a vector - valued function @xmath48 such that @xmath49 .
2 . @xmath50 lies on @xmath51 with @xmath52 . in the subsection [ hopf ] ,
the nonlinear perturbation @xmath53 is found directly related to the cole - hopf transformation .
+ we are ready to state the main results of the paper .
the first result is on the nonlinear stability of the planar traveling wave solution @xmath17 of the system when @xmath54 .
[ theoremnc ] for a given planar traveling wave solution @xmath55 of the system for @xmath54 with , and , there exist constants @xmath56 and @xmath57 such that if @xmath58 , then for any initial data @xmath59 in the form of @xmath60 and @xmath61 where @xmath62 and @xmath63 are @xmath64periodic in @xmath6 with @xmath65 , there exist a unique global classical solution @xmath66 of for @xmath54 in the form of @xmath67 where @xmath68 and @xmath69 are @xmath64periodic in @xmath6 for each @xmath70 and @xmath71 and @xmath72 .
moreover , @xmath73 satisfies the following inequality : @xmath74 1 .
the condition @xmath75 implies the zero integral condition on @xmath76 , @xmath77 .
equivalently , @xmath34 and @xmath78 satisfy @xmath79 which is preserved over time , provided it holds at @xmath80 by the mass conservation property of @xmath34 equation .
the weight function @xmath81 is of a same order as @xmath82 in @xmath83 direction .
more precisely , for a fixed traveling wave , there exists a constant @xmath84 satisfying @xmath85 which is proved in section @xmath86 .
theorem [ theoremnc ] implies the asymptotic convergence of the primary perturbation of @xmath87 to @xmath17 . from @xmath88
we get a weak asymptotical estimate @xmath89 in terms of @xmath90 and @xmath91 , we have @xmath92 4 . the one - sided decay appears naturally with respect to the solvability of @xmath93 in the infinite strip domain @xmath94 @xcite . , it is shown that for any @xmath76 satisfying @xmath95 there exist a vector function @xmath96 in @xmath97 such that @xmath93 in @xcite . ] indeed , if there is a vector function @xmath98 satisfying the divergence equation , assuming dirichlet or the periodic boundary condition on the lateral boundary , we have @xmath99 } { { \nabla}\cdot}{\varphi}dz dy = -\int_{[0 , { \lambda } ] } { \varphi}^1 ( x , y ) dy.\ ] ] define @xmath100 , then @xmath101 is found in @xmath102 because @xmath103 } { \varphi}^1 ( x , y ) dy \right)}^2 dx \le \int _ 0 ^{\infty } { \lambda}\int _ { [ 0 , { \lambda } ] } |{\varphi}^1|^2(x , y ) dy dx = { \lambda}\| { \varphi}^1\|_{l^2}^2.\ ] ] in this paper , we do not pursuit solving @xmath93 for a given perturbation @xmath76 .
instead we are given @xmath104 .
note that the condition @xmath105 implies @xmath106 , which is enough for @xmath76 to satisfy @xmath107 and to be consistent with the necessary condtion for solvability .
we refer to @xcite for the solvability of the divergence equation on the more general noncompact domain .
when the chemical diffusion is present @xmath13 , the above result of theorem [ theoremnc ] is hard to obtain for technical reasons .
instead , we obtain a linear stability of the planar traveling wave solution @xmath17 of the system when @xmath108 for perturbations in a smaller class(mean - zero in @xmath6 ) . indeed , as will be derived in section [ hopf ] , the pertubation @xmath109 satisfies @xmath110 where @xmath111 .
quadratic terms dropped out , the linearized system of @xmath112 is as follows : @xmath113 what it follows we denote @xmath114 .
[ theoremli ] for sufficiently small @xmath108 and for any given planar traveling wave solution @xmath55 of the system with , and , the wave @xmath55 is linearly stable in the following sense : + there exist constants @xmath115 and @xmath57 such that if @xmath58 , then for any initial data @xmath116 which is @xmath64periodic in @xmath6 with @xmath117 there exists a unique global classical solution @xmath112 of which is @xmath64periodic in @xmath6 and @xmath118 for each @xmath70 and which satisfies the inequality : @xmath119 the next theorem gives a hint why studying planar traveling waves is natural instead of studying general traveling waves in @xmath120 .
roughly speaking , any smooth bounded solution with our front boundary conditions , becomes eventually planar under some strong regularity conditions when @xmath108 .
this kind of nonlinear stability can be found in @xcite on the stability of a reactive boussinesq system in an infinite vertical strip
. it also hints the asymptotic stability as in theorem [ theoremnc ] might hold as well in a nonzero diffusion case .
we state this theorem in terms of @xmath121 in which will be obtained in the subsection [ hopf ] by applying the cole - hopf transformation @xmath122 to .
[ thm_eventual ] let @xmath108 and @xmath121 be a global smooth solution of it implies @xmath123 the remaining parts of the paper are organized as follows . in section [ backg ] , we introduce the backgroud materials including the existence of traveling wave solutions and the cole - hopf trasformation . in section [ sectionuniform ] ,
we prove proposition [ zerolocal ] and [ zeroglobal ] , then establish theorem [ theoremnc ] . in section [ sectionuniformwith ]
, we prove the energy inequality in theorem [ theoremli ] . with the existence part of theorem [ theoremli ]
follows the same argument that works for the @xmath54 case , therefore we omitted . in the last section ,
we proves theorem [ thm_eventual ] which points out a nonlinear stability on a thin cylindrical domain for @xmath108 case would be expected as well
. + there are many prior results on the existence of traveling wave solution and its stability . among earlier analytic works
is @xcite , where they proved the existence and linear instability of traveling wave solution of the one dimensional system when the bacterial consumption rate is constant and @xmath124 .
the existence of the traveling wave solution with the front conditions and was shown in @xcite when @xmath54 , and @xcite when @xmath108 . in the one dimensional system ,
the nonlinear asymptotic stability results were shown under the aforementioned restrictions on @xmath125 and @xmath126 in @xcite when @xmath54 , and @xcite when @xmath108 is small .
to our knowledge , our theorem [ theoremnc ] is the first result on a nonlinear stability of traveling wave solutions in a higher dimension .
+ for the results on the cauchy problem of , see @xcite , where @xcite prove the existence of a global weak solution , and @xcite proves that of a global classical solution .
both results consider the zero chemical diffusion case in a multi - dimension .
existence of traveling wave solutions[existence_traveling ] the traveling wave solution @xmath17 in solves the following ode system , @xmath127 it is easy to see the above system integrable .
we briefly discuss its solvability and properties of the solution for reader s convenience .
what it follows in this subsection has been already well known ( for instance in @xcite ) .
let @xmath128 .
we first solve @xmath78 by @xmath129 for some positive constant @xmath130 , which means a translation by @xmath131 when @xmath132 .
the system with the front conditions and is translation invariant .
it causes the undetermined constant @xmath130 . to satisfy the front condition
, we impose @xmath133 as @xmath134 , which implies @xmath135 is singular at zero such as @xmath136 .
substituting @xmath78 to the c equation , we have @xmath137 what it follows we let @xmath138 we result in @xmath139 by introducing @xmath140 we have a kpp - type equation for @xmath141 : @xmath142 since @xmath143 , front conditions , force @xmath141 to have @xmath144 it is well known that the kpp- fisher equation has a nonnegative solution if and only if @xmath145 , which automatically holds in this case .
more precisely we refer to the following lemma in @xcite , which is based on the standard argument for the fisher equation in @xcite etc .
[ kpptheorem][wang , lemma 3.2 . ] a nonnegative traveling wave solution @xmath141 of exists .
it satisfies @xmath146 @xmath147 moreover it is unique up to a translation in @xmath5 , or @xmath148 and has the following asymptotic behavior : @xmath149 with @xmath150 when @xmath108 and @xmath151 when @xmath54 . to be consistent with lemma [ kpptheorem ] , the wave speed @xmath152 is uniquely determined by @xmath153 , @xmath154 also from @xmath155 and the boundary condition @xmath156 , we have @xmath157 we convert back @xmath78 , @xmath158 satisfying from @xmath141 with the required front condition , and .
the monotonicity of @xmath159 follows from @xmath160 ; the profile of @xmath78 is decreasing , and @xmath158 is increasing left to rightward . , @xmath158 and @xmath161 . ] in the zero diffusion ( @xmath54 ) we can explicitly solve @xmath162 such that @xmath163 then @xmath164 we have the relation @xmath165 and the monotone profile of @xmath17 .
cole - hopf transformation[hopf ] by the cole - hopf transformation
@xmath166 we translate into the divergence type equations of @xmath121 @xmath167 with @xmath168 .
this works only for @xmath169 in .
the cole - hopf transformation is initiated in @xcite for the stability problem of the traveling wave equation in the one - dimensional angiogenesis .
it appears in @xcite studying the multi - dimensional angiogenesis equation for @xmath54 with vanishing boundary conditions , too .
+ we shall add the condition @xmath170 this curl free condition on @xmath171 is preserved in time until the classical local solution @xmath121 exists .
it holds that @xmath172 with .
we denote @xmath173 then the perturbation around @xmath17 is written by @xmath174 note that @xmath175 solves @xmath176 the boundary conditions of @xmath177 are inherited from those of @xmath17 : @xmath178 for a while we denote @xmath179 so the curl - free condition is endowed to @xmath180 .
the perturbation @xmath181 satisfies @xmath182 if we let @xmath183 the equation is put into the equation on @xmath109 @xmath184 from now on , we fix a planar traveling wave solution @xmath55 of the system for @xmath54 with , and , obtained in section [ existence_traveling ] .
the pertubation @xmath112 satisfies the system @xmath185 for @xmath186 $ ] where @xmath187 as in .
in addition , recall the function spaces and the periodic condition ( in @xmath6-direction ) for @xmath53 which are discussed prior to the main theorem [ theoremnc ] .
+ first we state a result on the local existence of solutions for the system which can be proved in the standard manner .
the proof is presented at the end of the section 3 ( subsection [ pf_local_exist ] ) .
[ zerolocal ] for @xmath1 and @xmath188 , there exists @xmath189 such that for any initial data @xmath190 which is @xmath64periodic in @xmath6 with @xmath191 , the system has a unique solution @xmath192 on @xmath193 $ ] which is @xmath64periodic in @xmath6 for each @xmath194 $ ] with @xmath71 and @xmath72 and @xmath195 moreover it holds the following inequality : @xmath196}\big ( \| \varphi\|^2_{h^3_w}+ \| \psi \|^2_{h^3 } + \| { \nabla}\psi\|^2_{h^2_w}\big ) \le 2 m.\ ] ] let us state the result on the existence of global classical solution to the system . due to our derivation of from
, the main theorem [ theoremnc ] is the direct consequence of proposition [ zeroglobal ] , which will be proved in section @xmath197 .
[ zeroglobal ] there exist constants @xmath198 , and @xmath57 such that for @xmath58 and for any initial data @xmath116 which is @xmath64periodic in @xmath6 with @xmath65 , there exists a unique global classical solution @xmath53 of which is @xmath64periodic in @xmath6 for each @xmath70 with @xmath71 and @xmath72 .
moreover , @xmath73 satisfies the inequality @xmath199 in the below we summarize the notations used in the paper . @xmath200 @xmath201 } ( \| \varphi(s)\|^2_{h^3_w } + \|\psi(s)\|^2_{h^3 } + \|{\nabla}\psi(s)\|^2_{h^2_w}),\ ] ] @xmath202 @xmath203
@xmath204 @xmath205 @xmath206 @xmath207
in this section we introduce the main proposition [ uniform _ ] below which implies proposition [ zeroglobal ] .
[ uniform _ ] there exist constants @xmath208 , and @xmath57 such that for @xmath209 $ ] , we have the following : + if @xmath109 be a local solution of on @xmath210 $ ] for some @xmath211 with @xmath212 for some initial data @xmath213 which is @xmath3-periodic in @xmath6 , then we have @xmath214 this proposition will be proved in subsection [ mainproof ] . note that the constant @xmath215 in the above proposition does not depend on the size of @xmath211 .
this fact immediately improves the bootstrap assumption if we take a sufficiently small initial data , for instance , satisfying @xmath216 .
once we prove the proposition , then the global - existence statement in proposition [ zeroglobal ] follows the standard continuation argument as shown below : let @xmath217 where @xmath218 and @xmath215 are the constants in proposition [ uniform _ ] .
we may assume @xmath219 .
consider the initial data @xmath116 with @xmath220 . by using this constant @xmath84 to the local - existence result ( proposition [ zerolocal ] ) ,
there exist @xmath189 and the unique local solution @xmath112 on @xmath221 $ ] with @xmath222 . due to @xmath223
, we can use the result of proposition [ uniform _ ] to obtain @xmath224 , which implies @xmath225 .
hence we can extend the solution up to the time @xmath226 by proposition [ zerolocal ] and we obtain @xmath227 . due to proposition [ uniform _ ] , it implies @xmath228 .
thus we can repeat this process of extension to get @xmath229 for any @xmath230 .
first we summarize some properties of traveling waves @xmath231 for later analysis .
properties on @xmath231 from the equations : @xmath232 we observe the relations @xmath233 from where @xmath234 , which is fixed throughout this paper , is the speed of our planar traveling wave @xmath55 .
in fact , @xmath177 has the explicit formula : @xmath235 note that @xmath78 and @xmath236 are monotone with the boundary condition . in the following lemma
, we summarize some properties of @xmath237 .
[ wave_prop ] there exists a constant @xmath188 such that we have : + @xmath238 @xmath239 @xmath240 ( @xmath241 is any @xmath242th derivative of @xmath78 ) the first inequality follows from lemma [ kpptheorem ] .
the others are easily obtained from the equation and .
proof of proposition [ uniform_][mainproof ] first assume @xmath243 . during the proof
, we will take @xmath218 as small as needed to have the weighted energy estimate ( e.g. lemma [ lemma1_2 ] ) .
let us remind the system @xmath244 for @xmath245\times { \mathbb r}\times [ 0 , { \lambda}]$ ] .
we will collect a few lemmas on energy estimates before proving proposition [ uniform _ ] . here
the constants @xmath158 which will appear in the following lemmas are independent of @xmath211 .
[ lemma0 _ ] if @xmath1 is sufficiently small , we have the following : @xmath246 this estimate is not closed due to the second term in the right - hand side .
this obstacle will be studied sequentially in the next lemmas .
multiply @xmath247 to the @xmath96 equation and @xmath50 to the @xmath50 equation . integrating by parts
, we have @xmath248 we estimate the cubic term first : @xmath249 where we used @xmath250 by the sobolev embedding .
thanks to the relations , the quadratic term becomes @xmath251 where @xmath252}{\varphi}(z , y)dy$ ] is the average in @xmath6 of @xmath68 for each @xmath38 .
note that we used the periodic condition for @xmath253 . as a result
, we get @xmath254 where we used the poincar inequality @xmath255 ) } \le c{\lambda}\| { \partial}_{y } { \varphi}(z,\cdot_y)\|_{l^2_y([0,{\lambda } ] ) } \quad\mbox { for } z\in\mathbb{r}.\ ] ] also we have @xmath256 thanks to the relations . by assuming smallness of @xmath3 , we get the following : @xmath257 it proves lemma [ lemma0 _ ] . + from now on we study the diffusion case(@xmath108 ) .
note that we do not have the simple relation .
instead , we estimate the quadratic term @xmath258 by @xmath259 where we used @xmath260 ( lemma [ wave_prop ] ) with the dirichlet condition for @xmath68 ( @xmath261 when @xmath262 or @xmath263 ) in order to use the poincar inequality @xmath264 ) } \le c{\lambda}\| { \partial}_{y } { \varphi}(z,\cdot_y)\|_{l^2_y([0,{\lambda } ] ) } \quad\mbox { for } z\in\mathbb{r}.\ ] ] also we estimate @xmath265 where we used the estimate @xmath266 in lemma [ wave_prop ] and the poincar inequality . for the @xmath267-terms
, we estimate @xmath268 for the first estimate , we used @xmath269 .
for the second estimate , we used the dirichlet condition for @xmath50 ( @xmath270 when @xmath262 or y=@xmath3 ) in order to use the poincar inequality @xmath271 ) } \le c{\lambda}\| { \partial}_{y } { \psi}(z,\cdot_y)\|_{l^2_y([0,{\lambda } ] ) } \quad\mbox { for } z\in\mathbb{r}.\ ] ] by assuming smallness of @xmath272 and @xmath3 , we arrive at @xmath273 which proves lemma [ lemma0 _ ] for @xmath108 . note that we used the smallness of @xmath267 to use lemma [ wave_prop ] .
let s study the first order estimate : [ lemma1_0 ] if @xmath1 is small enough , then we have @xmath274 this estimate is not closed yet due to the last two terms .
later in lemma [ lemma01 _ ] ( see the estimate ) , it will be closed by lemma [ lemma1_1 ] and lemma [ lemma1_2 ] .
indeed , the first one @xmath275 among these two will be covered by the second one @xmath276 ( see lemma [ lemma1_1 ] ) while the second term will be absorbed into the dissipation term @xmath277 on the left - hand side of ( see lemma [ lemma1_2 ] ) .
differentiating in @xmath38 , we have @xmath278 multiply @xmath279 and @xmath280 to the above equation . integrating by parts
, we have @xmath281 similarly , we get @xmath282 first , we estimate by using lemma [ wave_prop ] @xmath283 we estimate the quadratic terms as follows : @xmath284 @xmath285 the cubic terms are bounded by @xmath286 by @xmath287 and by the assumption @xmath288 . + for the @xmath267 terms , we estimate @xmath289 and @xmath290 adding up all the estimates above , we have @xmath291 after integration in time , thanks to lemma [ lemma0 _ ] , we can control the last term above so that we arrive at .
[ lemma1_1 ] if @xmath1 is small enough , then we have @xmath292 1 .
roughly speaking , this estimate moves @xmath293 to the denominator , while the integral multiplied by @xmath294 which can be assumed small enough .
this estimate for @xmath295 is due to the special structure .
note that the @xmath68-equation has the term @xmath296 on its right hand .
multiplying @xmath297 to the @xmath68-equation , we have @xmath298 let us rewrite the terms @xmath299 and @xmath300 .
@xmath301 and @xmath302 integration by parts gives @xmath303 by using the @xmath69-equation which has @xmath304 on its right side , we get @xmath305 integration on gives us @xmath306 we rearrange the above to get @xmath307 note that @xmath308 by @xmath309 .
for the second term @xmath310 , we estimate @xmath311 by bounding @xmath312 . for the @xmath267-term @xmath313 ,
we estimate @xmath314 and @xmath315 now integrating in time , we get @xmath316 by lemma [ lemma0 _ ] and by @xmath317 , we have .
what remains to close is to control the term @xmath318 .
note that @xmath50-equation lacks diffusion of closing the estimates without weight .
however , our choice of the one - sided exponential weight function @xmath319 gives a one - sided dissipation @xmath320 as shown below .
[ lemma1_2 ] if @xmath321 and @xmath1 are small enough , then we have @xmath322 first we take @xmath323 to the @xmath69-equation then multiply by @xmath324 to get @xmath325 note that for some @xmath326 @xmath327 integrating on each half strip ( notation : @xmath328 } f ( z , y , t)dy dz$ ] ) and in time , we get @xmath329 } w |{\nabla}{\psi}|^2(0 , y ) dy \quad ( \mbox{by } \eqref{positive_comp } ) \\\le & \int _ { z>0 } w |{\nabla}{\psi}_0|^2 -\frac c2 \int_0^t \int_{z>0 } w |{\nabla}{\psi}|^2 + c \int_0^t \int_{z>0 } w |{\nabla}({\nabla}\cdot { \varphi})|^2 \\ & -\frac s2\int_0^t \int_{[0,{\lambda } ] } w |{\nabla}{\psi}|^2(0 , y ) dy \end{aligned}\ ] ] and @xmath330 } w |{\nabla}{\psi}|^2(0 , y ) dy\\\le & \int_{z<0 } w |{\nabla}{\psi}_0|^2 + 2\int_0^t \int_{z<0 } |{\nabla}({{\nabla}\cdot}{\varphi})|| { \nabla}{\psi}| \\ & + \frac s2\int_0^t \int_{[0,{\lambda } ] } w |{\nabla}{\psi}|^2(0 , y ) dy \\
\le & \int_{z<0 } w |{\nabla}{\psi}_0|^2 + c\int_0^t \int_{z<0 } { \frac { |{\nabla}^2 { \varphi}|^2}{n}}+ \int_0^t \int_{z<0 } n |{\nabla}{\psi}|^2 \\ & + \frac s2\int_0^t \int_{[0,{\lambda } ] } w |{\nabla}{\psi}|^2(0 , y ) dy . \ ] ] adding the above two then adding @xmath331 to the both sides , we get @xmath332 where we used the previous estimate for the last inequality . for the @xmath267-terms , we estimate @xmath333 where we used the estimate @xmath334 and for the last inequality , we assumed @xmath272 small enough .
thus we have @xmath335 then , by making @xmath272 small enough , it proves the estimate .
we are ready to obtain a closed energy estimate up to the first order derivatives : [ lemma01 _ ] if @xmath321 and @xmath1 are small enough , then we have @xmath336 plugging the estimates and into , we have @xmath337 which gives us @xmath338 if we assume @xmath272 small enough .
in addition , from the estimate with the above estimate , we get @xmath339 adding lemma [ lemma0 _ ] and the above and and applying the estimate into the sum , we have .
note that we have used @xmath272 for @xmath340 , @xmath341 , and @xmath342 . from now on
, we will repeat the preceding energy estimates up to the third order derivatives to have the weighted energy estimate .
since the key ideas were already presented in detail on the previous lemmas , we do not split the higher order estimates into several lemmas as in the first order case ( lemma [ lemma0 _ ] , [ lemma1_0 ] , [ lemma1_1 ] , [ lemma1_2 ] and [ lemma01 _ ] ) but state them in the following single lemma , and we do not write its proof in detail .
[ lemma23_]if @xmath321 and @xmath1 are small enough , then for @xmath343 we get @xmath344 1 .
the proof for @xmath345 is similar with that of the first order estimate except and .
2 . for @xmath346
, we use @xmath347 ( see and ) as well as @xmath348 . differentiating the @xmath349 equations @xmath350 times in @xmath6 or @xmath38 , we have @xmath351 { \nabla}{\psi}+ n { \nabla}{{\partial}_y^{i}}{{\partial}_z^{j}}{\psi}+ { { \partial}_y^{i}}{{\partial}_z^{j } } ( p { { \nabla}\cdot}{\varphi } ) + { { \partial}_y^{i}}{{\partial}_z^{j } } ( { { \nabla}\cdot}{\varphi}{\nabla}{\psi}),\\ & { { \partial}_y^{i}}{{\partial}_z^{j}}{\psi}_t - s { { \partial}_y^{i}}{{\partial}_z^{j}}{\psi}_z = { { \nabla}\cdot}{{\partial}_y^{i}}{{\partial}_z^{j}}{\varphi}.\end{aligned}\ ] ] thus we get @xmath352 @xmath353 case @xmath354 + first , we estimate @xmath355 what it follows we do not distinguish @xmath356 and @xmath357 derivatives .
+ the quadratic terms are symbolically @xmath358 by lemma [ wave_prop ] , these terms are estimated by @xmath359 @xmath360 and @xmath361 where we used integration by parts for the last estimate .
the cubic terms are symbolically written as @xmath362 . by using integration by parts
once , it can be written as @xmath363 so by assuming smallness of @xmath272 , these terms are estimated by @xmath364 and @xmath365 note that we used @xmath272 for @xmath366 and @xmath342 . for the @xmath267 terms , we can write them symbolically : @xmath367 after intergration by parts , we can estimate them by @xmath368 and @xmath369 up to now , by lemma [ lemma01 _ ] , we estimate by @xmath370 this estimate is the second order version of lemma [ lemma1_0 ] .
+ now we claim the following two estimates which are the second order versions of lemma [ lemma1_1 ] and [ lemma1_2 ] : @xmath371 as we did in lemma [ lemma1_1 ] and [ lemma1_2 ] , our plan is to prove first and to use the result in order to get
. then we will close the estimate by using them .
+ @xmath353 proof of + taking @xmath372 to @xmath68 equation , we have @xmath373 we multiply @xmath374 on the both sides and plug it in the equation @xmath375 in order to get @xmath376 for @xmath377 , we get @xmath378
so , integrating on the strip , we have @xmath379 note that @xmath380 by @xmath381 . the terms in @xmath382 are estimated as follows ; @xmath383 by assuming smallness of @xmath272 . for the @xmath267 terms , we estimate them by @xmath384 for the second term and the third term
, we estimate @xmath385 and @xmath386 as a result , we get@xmath387 collecting the above estimates and using lemma [ lemma01 _ ] , we have @xmath388 to get from the above estimate , we have to control @xmath389 by @xmath390 with lower order estimates .
observe @xmath391 and @xmath392 by @xmath393 from .
thus we get @xmath394
so we have @xmath395 by lemma [ lemma01 _ ] .
thus we proved .
+ @xmath353 proof of + multiplying @xmath396 to the equation @xmath397 we have @xmath398 recall that for some @xmath326 , we have @xmath399 integrating on each half strip and in time , we get @xmath400 } w |{\nabla}^2 { \psi}|^2(0 , y ) dy \\ & \le \int _ { z>0 } w |{\nabla}^2 { \psi}_0|^2 -\frac c2 \int_0^t \int_{z>0 } w |{\nabla}^2 { \psi}|^2 + c \int_0^t \int_{z>0 } w |{\nabla}^2({\nabla}\cdot { \varphi})|^2 \\&\quad -\frac s2\int_0^t \int_{[0,{\lambda } ] } w |{\nabla}^2 { \psi}|^2(0 , y ) dy , \end{aligned}\ ] ] @xmath401 } w |{\nabla}^2 { \psi}|^2(0 , y ) dy \\ & \le \int_{z<0 } w |{\nabla}^2 { \psi}_0|^2 + c\int_0^t \int_{z<0 } | { \nabla}^3 { \varphi}|||{\nabla}^2 { \psi}| \\&\quad + \frac s2\int_0^t \int_{[0,{\lambda } ] } w |{\nabla}^2 { \psi}|^2(0 , y ) dy \\ & \le \int_{z<0 } w |{\nabla}^2 { \psi}_0|^2 + c\int_0^t \int_{z<0 } n|{\nabla}^2 { \psi}|^2 + c\int_0^t \int_{z<0 } { { \frac { |{\nabla}^3 { \varphi}|^2}{n } } } \\&\quad + \frac s2\int_0^t \int_{[0,{\lambda } ] } w |{\nabla}^2 { \psi}|^2(0 , y ) dy . \ ] ] adding those , we get @xmath402 for the @xmath267-terms , as before , we estimate @xmath403 where we used the estimate @xmath334 and for the last inequality , we assumed @xmath272 small enough .
+ collecting the above estimates , and using lemma [ lemma01 _ ] and the previous claim , we have @xmath404 then , by making @xmath272 small enough , it proves the claim .
+ now we are ready to finish this proof for lemma [ lemma23 _ ] for the second order . plugging into with @xmath272 small ,
we have @xmath405 in turns , we have @xmath406 this proves lemma [ lemma23 _ ] for case @xmath407 . together with [ lemma01 _ ] , we have proved @xmath408 + @xmath353 case @xmath409 + + we repeat the previous argument for @xmath346 .
first , we recall the equation . as before
, we estimate @xmath410 the quadratic terms are symbolically @xmath411 by lemma [ wave_prop ] , these terms are estimated by @xmath412 @xmath413 and @xmath414 where we used integration by parts for the last estimate . then by a similar argument
, we have @xmath415 the cubic terms are estimated by @xmath416 from @xmath417 ( lemma [ wave_prop ] ) , we can estimate each term .
indeed , recall @xmath418 so we get @xmath419 @xmath420 @xmath421 then the term @xmath422 remains .
all cubic terms can be estimated very similarly except @xmath422 .
note that by the sobolev embedding , @xmath423 so we estimate @xmath424 using @xmath425 ( lemma [ wave_prop ] ) , we get @xmath426 for the @xmath267-terms , we can write them symbolically : @xmath427 after integration by parts , we can estimate these terms by @xmath428 and @xmath429 collecting the above estimates and using , we get the third order version of : @xmath430 as before , we claim the following two estimates for the third order derivatives : @xmath431 these two claims can be proved similarly so we skip the proofs . we will prove below and we will use the result in order to get . then we will apply to close .
+ @xmath353 proof of + taking @xmath432 to @xmath68 equation where @xmath433 is either @xmath434 or @xmath435 , we have @xmath436 we multiply @xmath437 on both sides and plug in the equation @xmath438
so that @xmath439 for @xmath377 , we note @xmath440 so , integrating on the strip , we have @xmath441 note that @xmath442 by @xmath443 . + the integral containing @xmath382 is estimated as follows ; the quadratic terms : @xmath444 for small @xmath272 .
the last term in the above can be estimated : @xmath445 up to now , using , we have @xmath446 for @xmath267-terms , we estimate them by @xmath447 for the last two terms , we estimate @xmath448 and @xmath449 note that the last term @xmath450 is controlled by @xmath451 . as a result , we get @xmath452 we use again to get @xmath453 we replace @xmath433 with @xmath434 and @xmath435 and add these two estimates to get @xmath454 to get from the above estimate , we have to estimate @xmath455 from @xmath456 with lower order estimates .
we apply the estimate by replacing @xmath69 with @xmath297 then we get @xmath457 so @xmath458 by . together with , we proved .
+ @xmath353 proof of + multiplying @xmath459 to the equation @xmath460 we have @xmath461 as before , we will use @xmath462 where @xmath326 is a constant depending only on @xmath234 . integrating on each half strip and in time ,
we get @xmath463 } w |{\nabla}^3 { \psi}|^2(0 , y ) dy \\ & \le \int _ { z>0 } w |{\nabla}^3 { \psi}_0|^2 -\frac c2 \int_0^t \int_{z>0 } w
|{\nabla}^3 { \psi}|^2 + c \int_0^t \int_{z>0 } w |{\nabla}^3({\nabla}\cdot { \varphi})|^2 \\&\quad -\frac s2\int_0^t \int_{[0,{\lambda } ] } w |{\nabla}^3 { \psi}|^2(0 , y ) dy , \end{aligned}\ ] ] @xmath464 } w |{\nabla}^3 { \psi}|^2(0 , y ) dy \\ & \le \int_{z<0 } w |{\nabla}^3 { \psi}_0|^2 + c\int_0^t \int_{z<0 } | { \nabla}^4 { \varphi}|||{\nabla}^3 { \psi}| \\&\quad + \frac s2\int_0^t \int_{[0,{\lambda } ] } w |{\nabla}^3 { \psi}|^2(0 , y ) dy \\ & \le \int_{z<0 } w |{\nabla}^3 { \psi}_0|^2 + c\int_0^t \int_{z<0 } n|{\nabla}^3 { \psi}|^2 + c\int_0^t \int_{z<0 } { { \frac { |{\nabla}^4 { \varphi}|^2}{n } } } \\&\quad + \frac s2\int_0^t \int_{[0,{\lambda } ] } w |{\nabla}^3 { \psi}|^2(0 , y ) dy . \ ] ] adding the above two estimates , we get @xmath465 for @xmath267-terms , as before , we estimate @xmath466 where we used the estimate @xmath334 . for the last term(the cubic term ) , we estimate @xmath467 for small @xmath272 . + using the estimate and the previous claim , we obtain@xmath468 by making @xmath272 small enough
, we proved the claim .
+ now we can finish the proof of lemma [ lemma23 _ ] fully .
indeed plugging into with @xmath272 small , we have @xmath469 so from , we have @xmath470 this proves lemma [ lemma23 _ ] for case @xmath471 . finally , we obtain proposition [ uniform _ ] : adding lemma [ lemma01 _ ] and lemma [ lemma23 _ ] for @xmath472 , we have + @xmath473}\| { \varphi}(s)\|_{3,w}^2 + \| { \nabla}{\psi}(s)\|_{2,w}^2 + \| { \psi}(s)\|^2 } + \int_0^t \sum_{l = 1}^4 \| { \nabla}^{l } { \varphi}\|_w^2 + \int_0^t \sum_{l = 1}^3\| { \nabla}^l { \psi}\|_w^2 \\ & \le c { ( \| { \varphi}_0\|_{3,w}^2 + \| { \nabla}{\psi}_0\|_{2,w}^2 + \| { \psi}_0\|^2)}. \end{split}\ ] ] this proves proposition [ uniform _ ] .
note that the above estimate does not contain any control of @xmath474 .
this is not a problem when @xmath108 . indeed , for @xmath108
, thanks to the dirichlet condition ( in @xmath6 ) , we can use the poincar inequality ( see ) to get @xmath475 however , when @xmath54 , the periodic condition ( in @xmath6 ) is not enough to get such an estimate .
+ let us remind the equation : @xmath476 with data @xmath477 satisfying @xmath478 .
first we shall find the local solution @xmath479 of in @xmath480 @xmath481 for a sufficiently small @xmath482 .
we define the linear map @xmath483 mapping @xmath484 to @xmath485 .
@xmath486 we approximate the above linear system by galerkin method .
since @xmath487 is a separable hilbert space , there exists an orthonormal basis @xmath488 of @xmath489 .
we define @xmath490 by @xmath491 where @xmath492 solves the following ode system @xmath493 with @xmath494 here , @xmath495 is the vector valued function defined by @xmath496 where @xmath497 is the @xmath498 inner product of scalar valued functions , that is @xmath499 the initial data @xmath500 is determined by @xmath501 such that @xmath502 the determines the @xmath503 system of ordinary differential equations with respect to @xmath504 , hence @xmath505 exists globally . by multiplying @xmath506 on the left to and summing over @xmath507 , we have also @xmath508 ( the inner product of vectors are defined componentwise . ) + step 2 .
we apply the integration by parts to the term @xmath509 , and by usual energy estimates we arrive at @xmath510 with @xmath158 depending on @xmath511 . finally by gronwall s inequality
, we have the uniform estimate of @xmath512 in @xmath513 for a small @xmath482 depending only on @xmath514 and @xmath515 .
then @xmath512 converges weakly to @xmath516 , which is a distribution solution of .
note that is linear , so the weak convergence is sufficient for passing to the limit .
moreover @xmath517 satisfies @xmath518 for which we refer to theorem @xmath519 in chapter @xmath197 of @xcite . from the equation above , we have also @xmath520 .
in particular , @xmath517 is in @xmath521 .
we can write into the following duhamel s formula .
@xmath522 where @xmath523 .
the injectivity of the map @xmath524 is clear from the formula above .
we proceed as usual .
let @xmath525 and @xmath526 .
we can show by heat kernel estimates @xmath527 if @xmath482 is sufficiently small . by the fixed point theorem we obtain the solution of in @xmath513 .
then the similar energy estimates as presented in section 3 show that the solution @xmath109 satisfies @xmath528 and holds , for which we omit details .
in this section we introduce the proposition [ uniform_with ] below which implies theorem [ theoremli ] .
[ uniform_with ] there exist constants @xmath529 , and @xmath57 such that for @xmath530 , and @xmath209 $ ] , we have the following : + if @xmath109 be a local solution of on @xmath210 $ ] for some @xmath211 with @xmath531 for some initial data @xmath213 which is @xmath3-periodic in @xmath6 and which has the zero average in @xmath6 ( @xmath532 ) , then we have @xmath533 1 . for this linear case , we do not need any smallness of @xmath534 since such a condition can pop up only from the nonlinear term estimate which is not present in this linear case .
2 . as in proposition
[ uniform _ ] , @xmath215 does not depend on the size of @xmath211 . for the proof we shall proceed the similar line of arguments as for propotision [ uniform _ ] with a minimal modification .
first we recall that we assumed @xmath535 and @xmath536 .
since the coefficient functions in the system depend on @xmath38 only , the zero average conditions are preserved over time .
it enables us to use the poincar inequality to @xmath112 with respect to @xmath537variable : @xmath538 ) } \le c{\lambda}\| { \partial}_{y } { \varphi}(z,\cdot_y)\|_{l^2_y([0,{\lambda } ] ) } , \quad \| { \psi}(z,\cdot_y ) \|_{l^2_y([0,{\lambda } ] ) } \le c{\lambda}\| { \partial}_{y } { \psi}(z,\cdot_y)\|_{l^2_y([0,{\lambda } ] ) } , \ , z\in\mathbb{r}.\ ] ] second we collect some properties of traveling waves @xmath231 .
note that @xmath108 and @xmath539 solves @xmath540 with the boundary condition .
all the properties in lemma [ wave_prop ] hold still true when @xmath108 .
[ wave_propwith ] for @xmath541 , there exists a constant @xmath188 such that for any @xmath542 , we have : + @xmath238 @xmath239 @xmath240 ( @xmath241 is any @xmath242th derivative of @xmath78 ) the first inequality follows lemma [ kpptheorem ] . for the second line , it is obvious that @xmath543 case holds .
if @xmath544 @xmath545 are shown , the other items follow the equation . in
the below we present the proof of @xmath544 @xmath545 . + @xmath353 @xmath546 + this case @xmath547 was proved in @xcite .
we sketch its proof for readers convenience .
+ multiplying @xmath548 to the second equation of and integraing over @xmath549 , we have @xmath550 by using monotonicity of @xmath551 and @xmath552 .
next multiplying @xmath553 to the same equation , we can show @xmath554 on the other hand , from the first equation @xmath555 , we find @xmath556 thus we have @xmath557 also note that @xmath558 . plugging the bounds in the inequality @xmath559 we conclude @xmath560 .
+ @xmath353 @xmath345 + first we observe @xmath561 from the equations for any @xmath108 . multiplying @xmath562 to the equation @xmath563 , and integrating over @xmath549 ,
we have @xmath564 we use the monotonicity of @xmath551 and @xmath552 in the third line .
we find @xmath565 which can be shown bounded by substituting @xmath566 , @xmath567 with lower order terms following the equations and using @xmath568 , @xmath569 for @xmath570 and . on the other hand
, we find @xmath571 @xmath572 the first integral and the second term are bounded .
it remains to estimate @xmath573 and @xmath574 by the cauchy - schwartz inequality . from the first equation , it is easy to see @xmath575 . by multiplying @xmath548 to the equation @xmath563 and integrating over @xmath549 , we have @xmath576
so we have @xmath577 . finally we estimate @xmath578 plugging the estimates above into , we obtain @xmath579 .
proposition [ uniform_with ] is proved by the lemmas below which are parallel to lemma [ lemma0 _ ] , [ lemma1_0 ] , [ lemma1_1 ] , [ lemma1_2 ] , [ lemma01 _ ] , [ lemma23 _ ] . here
we present in detail the zeroth order estimate of @xmath108 case , where the zero average condition , hence the poincar inequality is essentially used . for the higher order estimates , we sketch their proof . [ lemma0_e ]
if @xmath1 is sufficiently small , we have the following : @xmath580 recall the equations .
multiply @xmath247 to the @xmath96 equation and @xmath50 to the @xmath50 equation .
integrating by parts , we have @xmath581 note that does not hold for @xmath108 .
instead , we estimate the quadratic term @xmath258 by @xmath259 where we used @xmath260 ( lemma [ wave_propwith ] ) with the mean - zero condition for @xmath62 ( @xmath582 ) , which is preserved in time , in order to use the poincar inequality @xmath264 ) } \le c{\lambda}\| { \partial}_{y } { \varphi}(z,\cdot_y)\|_{l^2_y([0,{\lambda } ] ) } \quad\mbox { for } z\in\mathbb{r}.\ ] ] also we estimate @xmath265 where we used the estimate @xmath266 in lemma [ wave_propwith ] and the poincar inequality . for the @xmath267-term
, we estimate @xmath583 where we used the mean - zero condition for @xmath50 to use the poincar inequality for @xmath50 : @xmath271 ) } \le c{\lambda}\| { \partial}_{y } { \psi}(z,\cdot_y)\|_{l^2_y([0,{\lambda } ] ) } \quad\mbox { for } z\in\mathbb{r}.\ ] ] by assuming @xmath1 small enough , we arrive at @xmath584 which proves lemma [ lemma0_e ]
. [ lemma1_0_e ] if @xmath1 is small enough , then we have @xmath585 differentiating in @xmath38 , we have @xmath586 multiply @xmath279 and @xmath280 to the above equation . integrating by parts
, we have @xmath587 similarly , we get @xmath588 we repeat the process in the proof of lemma [ lemma1_0 ] . for the @xmath267 term , we estimate @xmath589 the last term is absorbed into the chemical diffusion term
. then we have @xmath590 for the last two terms above we use lemma [ lemma0_e ] .
[ lemma1_1_e ] if @xmath1 is small enough , then we have @xmath591 as in the proof of lemma [ lemma1_1 ] , we get multiplying @xmath297 to the @xmath68-equation , we have @xmath592 let us rewrite the terms @xmath299 and @xmath377 . @xmath593 and @xmath594 integration by parts
gives @xmath595 by using the @xmath69-equation which have @xmath304 on its right side , we get @xmath596 integration on gives us @xmath597 we rearrange the above to get @xmath598 for @xmath599 and @xmath310 , we refer to the proof of lemma [ lemma1_1 ] .
note that @xmath308 by @xmath309 .
for the second term @xmath310 , we estimate @xmath600 for the @xmath267-term @xmath313 , we estimate @xmath601 and @xmath602 integrating in time , we get @xmath603 by lemma [ lemma0_e ] , we have . [ lemma1_2_e ]
if @xmath1 and @xmath108 are small enough , then we have @xmath604 here we used the smallness condition on @xmath267 first in order to get the weighted @xmath605-estimate for @xmath606 above as in lemma [ lemma1_2 ] .
first we take @xmath323 to the @xmath69-equation then multiply by @xmath324 to get @xmath607 note that for some @xmath326 @xmath399 integrating on each half strip and in time then summing these two , as in the proof of lemma [ lemma1_2 ] , we get @xmath608 where we used the previous estimate . for the @xmath267-terms , we estimate @xmath609 where we used the estimate @xmath334 .
plugging this estimate , we have @xmath610 then , by making @xmath267 small enough , it proves the estimate . by adding all the above lemmas , we get [ lemma01__e ] if @xmath1 and @xmath108 are small enough , then we have @xmath611 plugging the estimates into , we have @xmath612 in addition , from the estimate with the above estimate , we get @xmath613 adding lemma [ lemma0_e ] and the above and , we have
. then we can repeat this process up to the highest order : [ lemma23_e]if @xmath1 and @xmath614 are small enough , then for @xmath343 we get @xmath615 we skip the proof(refer to that of lemma [ lemma23 _ ] ) .
@xmath616 @xmath353 case @xmath354 + for the @xmath267 terms , after intergration by parts , we can estimate them by @xmath368 up to now , by lemma [ lemma01__e ] , we estimate by @xmath617 this estimate is the second order version of lemma [ lemma1_0_e ] .
+ now we claim the following two estimates which are the second order versions of lemma [ lemma1_1_e ] and [ lemma1_2_e ] : @xmath618 @xmath353 proof of + as we did in the proof of ( and lemma [ lemma1_1_e ] for @xmath267-control ) , we can get @xmath619 then by , we get .
+ @xmath353 proof of + as we did in the proof of ( and lemma [ lemma1_2_e ] for @xmath267-control ) , we can get @xmath620 then , by making @xmath267 , it proves the claim .
+ using and with , we have lemma [ lemma23_e ] for case @xmath407 . indeed together with lemma [ lemma01__e ]
, we have proved @xmath621 + @xmath353 case @xmath409 + + we repeat the previous argument for @xmath346 . collecting the above estiamtes and using
, we get the third order version of : @xmath622 as before , we can show the following two estimates for the third order derivatives : @xmath623 by using and with , we can prove lemma [ lemma23_e ] for case @xmath471 .
+ finally , we obtain proposition [ uniform_with ] : adding all the lemmas in this section , we have + @xmath624}\big(\| { \varphi}(s)\|_{3,w}^2 + \| { \nabla}{\psi}(s)\|_{2,w}^2 + \| { \psi}(s)\|^2}\big ) + \int_0^t \sum_{l = 1}^4 \| { \nabla}^{l } { \varphi}\|_w^2 + \int_0^t \sum_{l = 1}^3\| { \nabla}^l { \psi}\|_w^2 + { \epsilon}\int_0^t \| { \nabla}^{4 } { \psi}\|_w^2 \\ & \le c { ( \| { \varphi}_0\|_{3,w}^2 + \| { \nabla}{\psi}_0\|_{2,w}^2 + \| { \psi}_0\|^2)}. \end{split}\ ] ] note that the above estimate does not contain any control of @xmath474 .
this is not a problem when @xmath108 .
indeed , for @xmath108 , thanks to the mean - zero assumption ( in @xmath6 ) , we can use the poincar inequality ( see ) to get @xmath475 however , when @xmath54 , just having the periodic condition ( in @xmath6 ) is not enough to get such an estimate . taking a @xmath6-derivative in we have @xmath625 integration by parts
gives @xmath626 where @xmath627 is small parameter to be chosen .
the right hand side of the second inequality is bounded by @xmath628 we add and .
for sufficiently small @xmath627 , we make @xmath629 and @xmath630 terms of the right hand side smaller than @xmath631 . by the poincar inequality ( in @xmath6 ) we have @xmath632
then we have @xmath633 @xmath634 thanks to @xmath635 , we decrease @xmath1 to get @xmath636 thus , there exists a constant @xmath637 such that @xmath638
mc s work was supported by nrf-2015r1c1a2a01054919 .
kc s work was supported by the 2015 research fund(1.150064.01 ) of unist(ulsan national institute of science and technology ) and by nrf-2015r1d1a1a01058614 .
kk s work was supported by nrf-2014r1a2a1a11051161 and nrf-20151009350 .
jl s work was supported by nrf-2009 - 0083521 .
00 r. a. anderson and m.a , j .
chaplain , _ modelling the growth and form of capillary networks in on growth and form : spatio - tempotal pattern formation in biology _ , wiley , 225249 , 1999 .
brazhnik , j. j. tyson , _ on traveling wave solutions of fisher s equation in two spatial dimensions _ , siam j. appl .
60 , no 2 , 371391 , 2000 .
p. constantin , a. kiselev , and l. ryzhik , _ fronts in reactive convection : bounds , stability and instability _ , comm .
pure and appl .
, 56 , 17811803 , 2003 .
l. corrias , b. perthame , h. zaag , _ a chemotaxis model motivated by angiogenesis _ , c. r. math .
paris , 336(2 ) ; 141146 , 2003 .
l. corrias , b. perthame , h. zaag , _ global solutions of some chemotaxis and angiogenesis systems in high space dimensions _ , milan j. math . , 72(1 ) ; 128 , 2004 .
bruce alberts et al .
, _ molecular biology of the cell _ , 4th edition , garland science , new york , 2002
. m. a. fontelos , a. friedman and b. hu , _ mathematical analysis of a model for the initiation of angiogenesis , _ siam . j. math .
33 , 13301355 , 2002 .
a. friedman and j.i.tello , _ stability of solutions of chemotaxis equations in reinforced random walks _ , j. math .
272 , 136163 , 2002 .
jin , j. li and z.a .
wang , _ asymptotic stability of traveling waves of a chemotaxis model with singular sensitivity _
, j. differential equations , vol .
225 , 193210 , 2013 .
j.a.leach and d.j.needham .
_ matched asymptotic expansions in reaction - diffusion theory _
, springer monograghs in mathematics , 2004 .
h.a.levine , b. d. sleeman and m. nilson - hamilton . _ a mathemacial model for the roles of pericytes and macrophages in the onset of angiogenesis : i. the role of protease inhibitors in preventing angiogenesis . _ math .
biosci . , 168 , 77115 , 2000 .
d. li and t. li , _ on a hyperbolic - parabolic system modeling chemotaxis _ , math .
models methods appl .
21 , no . 8,16311650 , 2011 . j. li , t. li and z.a .
wang , _ stability of traveling waves of the keller - segel system with logarithmic sensitivity _
, math . models and methods in appl .
, vol 24 , no 14 , 28192849 , 2014 . t. li and z. a. wang , _ asymptotic nonlinear stability of traveling waves to conservation laws arising from chemotaxis _
, j. differential equations , vol .
250 , 13101333 , 2011 .
d. horstmann . _ from 1970 until present : the keller - segel model in chemotaxis and its consequences i _ , jahresber .
verein . , 105 , 103-165 , 2003 .
d. horstmann and a. stevens , _ a constructive approch to traveling waves in chemotaxis _ , j. nonlinear sci .
14 , 125 , 2004 .
e. f. keller and l. a. segel , _ traveling bands of chemotactic bacteria : a theoretical analysis _
30 , 235-248 , 1971 . j. d. murray mathematical biology i : an introduction .
springer , berlin , 3rd edition , 2002 .
t. nagai and t. ikeda , _ traveling waves in a chemotactic model _ , j. math .
30 , 169184 , 1981 . b. perthame .
pde models for chemotactic movement : parabolic , hyperbolic and kinetic .
_ applications of mathematics .
_ , 49 , 539564 , 2004 .
h. schwetlick , travelling waves for chemotaxis systems . proc
mech . , 2 : 476478 , 2003
. j. a. shettatt , travelling wave solutions of a mathematical model for tumor encapulation , siam j. appl .
2 , 392407 , 1999 . v. a. solonnikov , _ on the solvability of boundary and initial boundary value problems fir the navier - stokes system in domains with noncompact boundaries _
, pacific j. math .
2 , 443458 , 1981 .
r. temam , _ navier - stokes equations _ , ams chelsea publishing , 2001 .
z. a. wang , _ wavefront of an angiogenesis model _
, dcds - b , vol .
8 , 28492860 , 2012 .
z.a.wang and t. hillen , _ shock formation in a chemotaxis model _ , math .
methods , appl .
31 , 4570 , 2008 . | a simplified model of the tumor angiogenesis can be described by a keller - segel equation @xcite .
the stability of traveling waves for the one dimensional system has recently been known by @xcite . in this paper
we consider the equation on the two dimensional domain @xmath0 for a small parameter @xmath1 where @xmath2 is the circle of perimeter @xmath3 .
then the equation allows a planar traveling wave solution of invading types .
we establish the nonlinear stability of the traveling wave solution if the initial perturbation is sufficiently small in a weighted sobolev space without a chemical diffusion . when the diffusion is present , we show a linear stability .
lastly , we prove that any solution with our front conditions eventually becomes planar under certain regularity conditions .
the key ideas are to use the cole - hopf transformation and to apply the poincar inequality to handle with the two dimensional structure . |
in this paper , we derive estimates which compare the running maximum of a martingale with its quadratic variation . given real numbers @xmath1 , @xmath2 we write @xmath3_n:= x_0 ^ 2+\sum _ { k=0}^{n-1 } ( x_{k+1}-x_k)^2,\quad\quad ( h\cdot x)_n:= \sum_{k=0}^{n-1 } h_k(x_{k+1}-x_k ) .\ ] ] we will derive pathwise versions of the famous burkholder davis gundy inequalities .
[ bdgthm ] for @xmath4 , there exist constants @xmath5 such that the following holds : for every @xmath6 and every martingale @xmath7 @xmath8_n^{p/2}\leq a_p{\mathbb{e}}\bigl [ \bigl(x_n^*\bigr)^p \bigr ] , \quad\quad{\mathbb{e}}\bigl [ \bigl(x_n^*\bigr)^p \bigr ] \leq b_p { \mathbb{e}}[x]_n^{p/2}.\ ] ] for @xmath9 this was established by burkholder @xcite . under additional assumptions , burkholder and gundy @xcite obtain a version for @xmath10 $ ] , while the case @xmath11 of ( [ bdgineq ] ) without restrictions is due to davis @xcite . for a modern account
see , for instance , @xcite .
the novelty of this note is that the above martingale inequalities are established as consequences of _ deterministic _ counterparts .
we postpone the general statements and first state the _ trajectorial version _ of davis inequality .
[ pathdavisthm ] let @xmath12 be real numbers and set . ] @xmath13_n + ( x_n^*)^2 } } , n\leq n$ ] .
then @xmath14}_n \leq3 x^*_n - ( h\cdot x)_n,\quad\quad x^*_n \leq6 \sqrt{[x]}_n + 2 ( h\cdot x)_n.\ ] ] while the proof of theorem [ pathdavisthm ] is not trivial , we emphasize that the inequalities in ( [ pathdavis ] ) are completely elementary in nature .
the significance of the result lies in the fact that it implies davis inequalities : indeed , if @xmath15 is a martingale , we may apply ( [ pathdavis ] ) to each trajectory of @xmath16 and obtain a bounded and adapted process @xmath17 .
the decisive observation is that , by the martingale property , @xmath18= 0,\ ] ] so davis inequalities ( with @xmath19 ) follow from ( [ pathdavis ] ) by taking expectations .
we recall that the bdg inequalities also apply if @xmath20 is a cadlag local martingale , and that this follows from a straightforward limiting procedure .
moreover , the inequalities are considerably simpler to prove for _ continuous _ local martingales ( see , for example , @xcite ) ; in this case , they also hold for @xmath21 , as proved by burkholder and gundy @xcite .
the problem of finding the optimal values of the constants @xmath22 is delicate , and has been open for 47 years and counting ; we refer to osekowski @xcite for a discussion of the current state of research .
the rest of this paper is organized as follows . in section [ history ] , we discuss the history of the pathwise approach to martingale inequalities . in section
[ heuristics for the pathwise hedging approach ] , we explain the intuition behind the hedging strategy @xmath23 used in the pathwise version of davis inequality . in section [ davis inequality for continuous local martingales ] , we give a short proof of one davis inequality for continuous martingales ; notably , this argument leads to a better constant compared to the previous literature ( to the best of our knowledge ) . in section [ davis inequality ] , we establish theorem [ pathdavisthm ] . in section [ pathwise burkholder - gundy inequality ] , we use theorem [ pathdavisthm ] to derive trajectorial versions of the bdg - inequalities in the @xmath24 case ; these also lead to their corresponding classical probabilistic counterpart , thus concluding a fully analytic derivation of theorem [ bdgthm ] .
the inspiration of the pathwise approach to martingale inequalities used in this paper comes from mathematical finance , more specifically , the theory of model - independent pricing .
the starting point of the field is the paper @xcite of hobson , which introduces the idea to study option - prices by means of _ semi - static hedging _ ; we explain the concepts using the inequality @xmath25}_n
\leq3 x^*_n - ( h\cdot x)_n\ ] ] appearing in theorem [ pathdavisthm ] .
if the process @xmath26 describes the price evolution of a financial asset , the functions @xmath27}_n$ ] and @xmath28 have the natural financial interpretation of being exotic options ; specifically , here @xmath29 is an option on realized variance , while @xmath30 is a look - back option .
the seller of the option @xmath29 pays the buyer the amount @xmath31 after the option s expiration at time @xmath32 , and @xmath33 corresponds to the gains or losses accumulated while trading in @xmath34 according to the portfolio @xmath23 .
the decisive observation of hobson is that inequalities of the type ( [ semistatic ] ) can be used to derive _ robust bounds _ on the relation of the prices of @xmath29 and @xmath30 : independently of the market model , one should never trade the option @xmath29 at a price higher than the price of @xmath30 , since the payoff @xmath29 can be _ super - hedged _ using the option @xmath30 plus self - financing trading . here
the _ hedge _
@xmath35 is designated _ semi - static _ : it is made up of a static part
the option @xmath36 which is purchased at time @xmath37 and kept during the entire time range plus a dynamic part which corresponds to the trading in the underlying asset according to the strategy @xmath38 .
since the publication of @xcite a considerable amount of literature on the topic has evolved ( e.g. , @xcite ) ; we refer in particular to the survey by hobson @xcite for a very readable introduction to this area .
the most important tool in model - independent finance is the skorokhod - embedding approach ; an extensive overview is given by obj in @xcite . starting with the papers @xcite the field has also been linked to the theory of optimal transport , leading to a formal development of the connection between martingale theory and robust hedging ( @xcite ) .
a benefit for the theory of martingale inequalities is the following guiding principle : _ every martingale inequality which compares expectations of two functionals has a deterministic counterpart . _ in fact , recently bouchard and nutz @xcite have coined this into a rigorous theorem in the discrete time setup , see also @xcite .
this idea served as a motivation to derive the doob - maximal inequalities from deterministic , discrete - time inequalities in @xcite ..
2pt in the present article , we aim to extend the approach to the case of the burkholder
davis gundy inequalities .
the aim of this section is to explain the basic intuition which lies behind the choice of the integrand in the pathwise davis inequalities .
arguments are simpler in the case of brownian motion , which we will now consider .
we focus on one of the two inequalities ; according to the pathwise hedging approach , we should be looking for a strategy @xmath17 and a constant @xmath39 such that @xmath40 . indeed , a reasonable ansatz to find a super hedging strategy is to search for a function @xmath41 such that @xmath42 to make an educated guess for the function @xmath43 , we argue on a _ purely heuristic _ level and consider paths which evolve in a very particular way .
assume first that the path @xmath44 stays infinitesimally close to the value @xmath45 for all @xmath46 : we picture bm as a random walk on a time grid with size @xmath47 , making alternating up and down steps of height @xmath48 .
thus , we assume that @xmath49 evolves in the form @xmath50 where necessarily @xmath45 lies between @xmath51 and @xmath52 .
the left - hand side of ( [ bdg1ansatz ] ) is of course increasing , so we have to ensure the same behavior on the right side .
a little calculation reveals that this means that @xmath43 should have the form @xmath53 to see this , set @xmath54 and compare the value @xmath55 with @xmath56 \sqrt{\mathrm{d}t } + \mathrm{o}\bigl(\mathrm{d}t^{3/2}\bigr ) \\ & \approx&- f_b \,\mathrm{d}t.\end{aligned}\ ] ] to assure that both sides of ( [ bdg1ansatz ] ) grow at the same speed , we thus need to require @xmath57 which leads to ( [ heu1 ] ) .
next , we consider a path which exhibits a different kind of extreme evolution : assume that @xmath58 for some number @xmath59
. simply setting @xmath60 would lead to @xmath61 .
taking @xmath62 sufficiently large , this quantity would eventually supersede @xmath63 independent of the choice of @xmath39 , and thus ( [ bdg1ansatz ] ) would fail .
so , this argument suggest to choose a function which is bounded ( at least for fixed @xmath64 ) .
moreover , dealing with a bounded integrand would conveniently allow to follow the explanation after theorem [ pathdavis ] and obtain davis inequalities from the pathwise davis inequalities .
thus , we could consider the function @xmath65 thanks to the additional term @xmath66 in ( [ bdg1ansatz ] ) , it is not a problem if @xmath67 is violated for `` small '' values of @xmath62 ; and , if @xmath68 is large compared to @xmath69 , @xmath70 holds , thus satisfying ( [ heu1 ] ) .
another similar possibility would be to use the function @xmath71 as in theorem [ pathdavisthm ] ; the latter turns out to lead to easier computations in the discrete time case .
we choose however @xmath43 given by ( [ contstrat ] ) when dealing with continuous martingales , since this allows us to obtain davis inequality with a better constant than the values we could find in the literature .
we now derive one pathwise davis inequality for continuous local martingales ; integrating it yields the corresponding standard davis inequality .
we notice that theorem [ bdg1bt ] provides the constant @xmath72 , which is smaller than the optimal constant for general cadlag martingales ( which is known to be @xmath73 , see @xcite ) .
we do not address here the opposite pathwise davis inequality for continuous local martingales since we are only interested in theorem [ thm4.1 ] for illustrative purposes ( since , as mentioned before , davis inequality for cadlag local martingales follows from the case of martingales in discrete time ) .
[ thm4.1 ] if @xmath74 is a continuous local martingale such that @xmath75 then @xmath76_t } \leq\frac{3}{2 } m_t^*- \biggl(\frac{m_t}{\sqrt { [ m]_t } \vee m^*_t}\cdot m_t \biggr)_t \quad\quad\mbox{for all } t\geq0 .\ ] ] by the dambis dubins schwarz time change result , it is enough to consider the case where @xmath74 is a brownian motion , which we will denote by @xmath49 . from ito s
formula applied to the semi - martingales @xmath77 and @xmath78 we find @xmath79 we may thus replace the integral in ( [ bdg1bt ] ) and arrive at the equivalent formulation @xmath80 inequality ( [ bdg1tbe ] ) gets stronger if we replace each occurrence of @xmath49 by @xmath69 ; thus , setting @xmath81 , it is enough to prove the following claim : let @xmath82 be continuous increasing functions such that @xmath83 and @xmath84 if @xmath85 .
then , for all @xmath85 @xmath86 to show this , observe that , by a change of variables @xmath87 .
hence , integrating by parts on the interval @xmath88 and taking the limit @xmath89 , we see that the left hand side of ( [ techcore ] ) equals @xmath90 by a change of variables and applying trivial inequalities we obtain @xmath91 if @xmath92 , the last integral equals @xmath93 ; otherwise there exists some @xmath94 such that @xmath95 , and then evaluating separately the integral on @xmath96 and on @xmath97 we obtain that @xmath98 since @xmath99 holds for @xmath100 , either way ( [ techcore ] ) follows .
in this section , we prove theorem [ pathdavisthm ] ; in fact , we will establish that ) slightly improves on inequality ( [ pathdavis ] ) by replacing the constant @xmath101 with the smaller @xmath102 . ]
@xmath103_n } & \leq&(\sqrt{2}+1 ) x^*_n+(-h \cdot x)_n , \\
\label{2half } x^*_n & \leq&6 \sqrt{[x]_n}+(2h \cdot x)_n , \end{aligned}\ ] ] where the dynamic hedging strategy is defined by @xmath104_n + ( x_n^*)^2}}$ ] as in theorem [ pathdavisthm ] . to prove ( [ 1half ] ) , ( [ 2half ] ) we introduce the convention , used throughout the paper , that any sequence @xmath105 is defined to be @xmath37 at time @xmath106 , and we define the auxiliary functions @xmath107 for @xmath108 , @xmath109 by @xmath110 and continuously extend them to @xmath111 by setting @xmath112 .
we will need the following lemma , whose proof is a somewhat tedious exercise in calculus .
[ davisineqfn ] for @xmath113 we have , with @xmath114 , @xmath115 \label{importantineq1 } g\bigl(x+d , m \vee|x+d| , q+d^2\bigr ) - g(x , m , q ) & \leq&- \frac{xd}{\sqrt { m^2+q}}+ c \bigl ( \bigl(m\vee|x+d|\bigr ) -m \bigr).\quad\quad\end{aligned}\ ] ] before proving lemma [ davisineqfn ] we explain why it implies ( [ 1half ] ) and ( [ 2half ] ) . proof of theorem [ pathdavisthm ] since @xmath116 , ( [ importantineq2 ] ) gives @xmath117_n } + x_n^*/2 & \leq & f\bigl(x_n , x_n^ * , [ x]_n\bigr ) \leq\sum_{k=0}^{n-1 } f\bigl(x_{k+1 } , x^*_{k+1 } , [ x]_{k+1}\bigr)-f \bigl(x_k , x_k^ * , [ x]_k\bigr ) \\[-0.5pt ] & \leq&(h\cdot x)_n + \sqrt{[x]_n},\end{aligned}\ ] ] which implies ( [ 1half ] ) ; and since @xmath118 , we get ( [ 2half ] ) from ( [ importantineq1 ] ) as follows @xmath119_n } & \leq & g\bigl(x_n , x_n^ * , [ x]_n\bigr ) \leq\sum_{k=0}^{n-1 } g\bigl(x_{k+1 } , x^*_{k+1 } , [ x]_{k+1}\bigr)-g \bigl(x_k , x_k^ * , [ x]_k\bigr ) \\[-0.5pt ] & \leq&-(h\cdot x)_n + c x^*_n.\vadjust{\goodbreak}\end{aligned}\ ] ] now we prove lemma [ davisineqfn ] .
proof of inequality ( [ importantineq1 ] ) it is enough to consider the case @xmath120 , as the one where @xmath121 then follows by continuity .
then , we can assume that @xmath122 through normalization .
define @xmath123 to be the lhs minus the rhs of ( [ importantineq1 ] ) ; since @xmath124 , it is sufficient to deal with the case @xmath125 .
_ case _ i @xmath126 $ ] : here we have to show that @xmath127 since @xmath128 , @xmath38 is convex , so it is sufficient to treat the boundary cases @xmath129 and @xmath130 . to simplify notation ,
we set @xmath131 ; notice that @xmath132 and @xmath133 .
_ sub - case _
i.a @xmath134 $ ] : then ( [ d1 ] ) follows from @xmath135 which is true since @xmath133 . _ sub - case _
i.b @xmath136 $ ] : here ( [ d1 ] ) amounts to @xmath137 _ case _ ii @xmath138 $ ] : since @xmath139 and @xmath125 , we find that @xmath140 implies @xmath141 . in this case @xmath38 equals @xmath142 since @xmath143 is convex , @xmath144 holds iff it holds for all @xmath34 on the boundary . moreover if @xmath145 , then we already know that @xmath144 from the corresponding sub - case @xmath146 and for @xmath147 , respectively . _ sub - case _ ii.a [ @xmath148 : we have to show that , for all @xmath149 , @xmath150 since @xmath151 and @xmath152 is convex , it follows that @xmath153 is convex in @xmath154 ; hence , the inequality has to be checked only for @xmath155 and for @xmath156 .
the first case is trivial , and in the latter , after dividing both sides by @xmath154 , we arrive at @xmath157 , which holds by our choice of @xmath158 and the fact that @xmath159 . _ sub - case _ ii.b [ @xmath160 : we have to show that , for all @xmath161 , @xmath162 as above , by convexity in @xmath154 it suffices to consider the cases @xmath163 and @xmath156 .
the first one amounts to @xmath164 , which is easily proved taking the squares .
the second one , after dividing by @xmath154 , amounts to @xmath165 , which holds since @xmath166 by our choice of @xmath158 . proof of inequality ( [ importantineq2 ] ) as before , we can assume w.l.o.g . that @xmath122 and @xmath125 .
define @xmath167 to be the lhs minus the rhs of ( [ importantineq2 ] ) .
_ case _ i [ @xmath168 : in this case , @xmath169 equals @xmath170 let us first isolate the terms that depend on @xmath34 .
define @xmath171 , and @xmath172 , so that @xmath173 notice that we can write @xmath174 and similarly @xmath175 for @xmath176 since the @xmath177 do not depend on @xmath34 and @xmath178 , @xmath179 . since @xmath180 equals @xmath37 if @xmath181 and equals @xmath182 if @xmath183 , to show @xmath184 we are lead to study the following two sub - cases .
_ sub - case _ i.a [ @xmath185 : in this case , @xmath186 ; to show that @xmath187 it is enough to show @xmath188 .
since @xmath189 we get @xmath190 , and so trivially @xmath191 so , calling @xmath192 , it is enough to prove that for all @xmath193 @xmath194 which is seen to be true by taking squares and bringing everything on the rhs to obtain a polynomial whose coefficients are all positive .
_ sub - case _ i.b [ @xmath195 : in this case @xmath196 , so it is enough to show that @xmath197 . since from @xmath198
it follows that @xmath199 , computations entirely similar implies @xmath200 . ] to the other sub - case establish the desired result .
_ case _ ii [ @xmath201 : in this case , @xmath202 and @xmath169 equals @xmath203 since trivially @xmath204 , to show @xmath184 we can assume that @xmath130 , in which case we can write @xmath169 as @xmath205 for @xmath206 since @xmath207 $ ] we have @xmath208 , and so to get @xmath184 it suffices to show that @xmath209 for @xmath210 .
this holds since @xmath211
garsia has given a simple proof of the fact that the bdg inequalities for general @xmath212 are a consequence of davis inequality ( @xmath11 ) and of the famous lemma by garsia and neveu ; in this section we revisit his proof and turn it into pathwise discrete - time arguments .
garsia s proof ( for which we refer to @xcite , chapter 3 , theorems 30 and 32 or to @xcite ) works similarly to how the doob @xmath213-inequalities for @xmath24 follow by writing @xmath214 as an integral , applying the ( weak ) doob @xmath215-inequality , using fubini s theorem , and finally applying h " older s inequality ( see for example @xcite ) .
the difference is that for the bdg inequalities one needs to use a different integral expression for @xmath214 , and so one has to consider davis inequalities not on the time interval @xmath216 $ ] but on @xmath217 $ ] , where @xmath218 is a stopping time . in the pathwise
setting , by the guiding principle stated in section [ history ] , if @xmath219 is a functional of a martingale @xmath16 and @xmath218 is a stopping time , a statement of the type @xmath220\leq0 $ ] will have to be turned into one of the type @xmath221 ; moreover , since there will be no expectations involved , hlder s inequality will have to be replaced by young s inequality .
we will need to consider discrete time stochastic integrals for which the initial time is different from @xmath37 ; given @xmath222 and real numbers @xmath223 and @xmath224 , we define @xmath225 moreover if , for @xmath226 , @xmath227 is a _ function _ from @xmath228 to @xmath229 , given real numbers @xmath230 we define @xmath231 as @xmath232 either way , we set @xmath233 if @xmath234 .
we now deduce pathwise davis inequalities on @xmath235 from the ones on @xmath236 by a simple time shift .
[ conddavis ] assume that @xmath237 and @xmath238 satisfy @xmath239_n } \leq\alpha x^*_n+(h \cdot x)_n,\quad\quad x^*_n \leq\beta \sqrt{[x]_n } + ( k\cdot x)_n\ ] ] for every sequence @xmath240 .
define , for @xmath241 , @xmath242 , the functions @xmath243 by @xmath244 then we have , for @xmath245 , @xmath246_n}-\sqrt{[x]_{i-1 } } \leq2 \alpha x^*_n + \bigl(f^{(i)}\cdot x\bigr)^n_i,\quad\quad x^*_n -x^*_{i-1}\leq\beta\sqrt{[x]_n } + \bigl(g^{(i)}\cdot x\bigr)^n_i.\ ] ] fix @xmath245 , @xmath240 and let @xmath247 . applying ( [ 1dav ] ) to @xmath248 we find @xmath249_n}-\sqrt{[x]_{i-1}}&\leq&\sqrt{[x]_n-[x]_{i-1 } } = \sqrt { \bigl[y^{(i)}\bigr]_{n - i } } \leq \alpha \bigl(y^{(i)}\bigr)^*_{n - i } + \bigl(h \cdot y^{(i ) } \bigr)_{n - i } \\ & \leq&\alpha2 x^*_n+\bigl(f^{(i ) } \cdot x \bigr)_{i}^n,\end{aligned}\ ] ] and ( respectively ) @xmath250_{n - i}}+\bigl(k \cdot y^{(i)}\bigr)_{n - i } \leq \beta\sqrt { [ x]_n}+\bigl(g^{(i ) } \cdot x\bigr)_{i}^n .\ ] ] here follows the pathwise version of garsia
neveu s lemma .
[ garsianeveu ] let @xmath24 , @xmath251 , and assume that @xmath252 and @xmath253 then , if we set @xmath254 we have that @xmath255 from @xmath256 and @xmath257 on @xmath258 $ ] , we find ( [ gnlc1 ] ) by writing @xmath259 \\ & = & pc_n a_{n}^{p-1 } + p \sum _ { i=0}^{n } \sum_{j = i}^{n-1 } \bigl(a^{p-1}_i - a^{p-1}_{i-1}\bigr ) h_j^{(i ) } ( x_{j+1}-x_j ) \\
& = & pc_n a_{n}^{p-1 } + \sum _ { j=0}^{n-1 } \biggl ( \sum_{i=0}^{j } p \bigl(a^{p-1}_i - a^{p-1}_{i-1}\bigr ) h_j^{(i ) } \biggr ) ( x_{j+1}-x_j ) = pc_n a_{n}^{p-1 } + ( w\cdot x)_n.\end{aligned}\ ] ] we then obtain ( [ gnlc2 ] ) from ( [ gnlc1 ] ) by applying young s inequality @xmath260 ( where @xmath261 and @xmath262 ) with @xmath263 , @xmath264 .
finally , from theorem [ pathdavisthm ] , lemma [ conddavis ] and lemma [ garsianeveu ] , we obtain the following discrete - time pathwise bdg inequalities for @xmath24 .
we recall that , by convention , @xmath265_{-1}=0 $ ] and @xmath266 , and in particular the integrand @xmath267 is well defined .
[ pathbgthm ] let @xmath12 be real numbers , @xmath268 for @xmath24 , and define @xmath269_i^{p-1 } } -\sqrt { [ x]_{i-1}^{p-1 } } \bigr ) f^{(i)}_n,\quad\quad g_n:= \sum_{i=0}^n p^2 \bigl(\bigl(x^*_i\bigr)^{p-1 } - \bigl(x^*_{i-1}\bigr)^{p-1 } \bigr ) f^{(i)}_n,\ ] ] where @xmath270_n-[x]_{i-1 } + \max_{i\leq k\leq n}(x_{k}-x_{i-1})^2}}.\ ] ] then @xmath271_n^{p } } \leq c_p \bigl(x^*_n\bigr)^p - ( h\cdot x)_n,\quad\quad \bigl(x^*_n\bigr)^p \leq c_p \sqrt{[x]_n^{p } } + 2(g\cdot x)_n.\ ] ] we notice that theorem [ pathbgthm ] yields ( [ bdgineq ] ) ; indeed , given a finite constant @xmath32 and a martingale @xmath15 , trivially @xmath272_n}$ ] and @xmath273 are in @xmath274 iff @xmath275 is in @xmath274 for every @xmath276 , and in this case the adapted integrands @xmath277 and @xmath278 which we obtain applying theorem [ pathbgthm ] to the paths of @xmath16 are in @xmath279 for every @xmath280 ( for @xmath281 ) , thus @xmath282 and @xmath283 are martingales and so @xmath284= 0= { \mathbb{e}}\bigl [ ( g\cdot x)_n\bigr],\ ] ] and the burkholder davis
gundy inequalities for @xmath24 ( with @xmath285 ) follow from ( [ pathbg ] ) by taking expectations , completing the proof of theorem [ bdgthm ] .
the authors thank harald oberhauser for comments on an earlier version of this paper .
the first author thanks the austrian science fund for support through project p21209 . | we present a new proof of the burkholder davis gundy inequalities for @xmath0 .
the novelty of our method is that these martingale inequalities are obtained as consequences of elementary _ deterministic _ counterparts .
the latter have a natural interpretation in terms of robust hedging . |
heating by supernovae and active galactic nuclei ( agn ) are thought to play important roles in galaxy formation , generating the observed correlation between the bulge mass of a galaxy and the mass of the central supermassive black hole , the upper mass cutoff of galaxies and reheating the gas in galaxies , groups and clusters . in order to reconcile the predictions of the cold dark
matter hierarchical clustering scenario with the observed luminosity function of galaxies and their mass - to - light ratios , it is thought that supernova feedback regulates star formation in galaxies less luminous than about @xmath4 , while agn feedback regulates star formation in more massive galaxies ( e.g. , white & frank 1991 ; bower et al .
2005 ; croton et al .
only type ii supernovae provide a natural feedback mechanism due to the short time delay between the onset of star formation and the first supernova . while type ia supernovae ( sn ia ) can not provide a feedback mechanism that regulates concurrent star formation , sn ia can play a significant role in the fate of gas shed by evolving stars in low - mass galaxies long after star formation ceases .
einstein observations showed that the bulk of the x - ray emission from optically luminous early - type galaxies arises from hot gas in hydrostatic equilibrium ( forman , jones & tucker 1985 ) .
a harder x - ray component , more prevalent among low - luminosity early - type galaxies , was detected by asca and assumed to arise from low - mass x - ray binaries ( lmxbs ) due to the old stellar population in these systems ( kim , fabbiano & trinchieri 1992 ) .
based on the analysis of asca data , white , sarazin & kulkarni ( 2002 ) found that the luminosity of the hard x - ray component was more strongly correlated with the luminosity in globular clusters rather than the total optical luminosity of the galaxies .
chandra , with its superior angular resolution , has resolved populations of point sources in many early - type galaxies ( e.g. , sarazin , irwin & bregman 2000 ; angelini , lowenstein & mushotzky 2001 ; kraft et al . 2001
; blanton , sarazin & irwin 2001 ) and found that 20 to 80% of the point sources reside in globular clusters ( sarazin et al .
2003 ) .
while the hot gas content of x - ray luminous early - type galaxies has been well studied by chandra and previous x - ray telescopes , little is known about the properties of the gas in less luminous early - type galaxies . in low - luminosity early - type galaxies ,
the x - ray emission from lmxbs can exceed the thermal emission from hot gas , so it is imperative that the lmxbs be detected and excised from the analysis of the diffuse emission .
theoretical models concerning the evolution of early - type galaxies show that for typical sn ia rates , early - type galaxies fainter than @xmath5 should possess sn ia driven galactic winds ( e.g. , david , forman & jones 1990 , 1991 ) .
in addition to heating by sn ia , agn heating may also expel a significant portion of the gas shed by evolving stars in the shallow potential well of low - luminosity galaxies .
chandra has detected x - ray cavities filled with radio emitting plasma and agn driven shocks in many elliptical galaxies and clusters ( e.g. , mcnamara [ cols="<,^,^,^,^,^,^,^ " , ] notes : galaxy name , 1.4 ghz flux ( @xmath6 ) and luminosity ( @xmath7 ) of the central agn . hole mass and the absolute red magnitude , @xmath8 , found by mclure & dunlop ( 2002 ) of @xmath9 , and an average @xmath10 of 2.5 for the galaxies in our sample , we plot the expected agn and sn ia heating rates in fig . 11 against the bolometric luminosity of the hot gas . for comparison , the mechanical power required to generate the observed shocks in ngc4552 ( the upper right hand point in fig .
11 ) is @xmath11 erg s@xmath12 ( machacek et al .
2005 ) , which is a factor of 2 greater than the rate predicted by best et al .
( 2006 ) , but certainly within the uncertainties .
11 indicates that heating by sn ia should dominate the energetics of the gas in early - type galaxies with @xmath13 and that agn heating should dominate in more luminous galaxies .
if heating by sn ia is the dominant mechanism for expelling the gas shed by evolving stars from low - luminosity early - type galaxies , it is a puzzle as to why there is so much scatter in @xmath14 and @xmath15 ( see figs . 7 and 11 ) given that supernova heating is essentially a continuous process .
since the gas masses are computed in different apertures ( see table 5 ) , we checked to see if there is a correlation between @xmath14 and aperture size and found none .
while the time - averaged sn ia heating rate shown in fig .
11 exceeds the time - averaged agn heating rate , except for possibly the most luminous galaxies in our sample , the episodic nature of agn outbursts could produce times during which mechanical agn heating exceeds sn ia heating .
thus , the large observed scatter could be a reflection of recent agn activity .
the sound crossing time within the central 10 kpc in these galaxies is approximately @xmath16 yr .
if agn outbursts repeat on a shorter time scale , then the gas will not be able to establish a steady - state wind .
in the very center of the galaxies , heating by sn ia can not be thought of as a continuous process .
for example , within an enclosed stellar mass of @xmath17 , there will only be one sn ia every @xmath18 yr . during this time
, the stars within this region will shed a total of 50 @xmath19 of gas .
if this gas is converted into mechanical energy by the central black hole with an efficiency of 10% , this will produce @xmath20 erg of energy at a rate of @xmath21 erg s@xmath12 , which is comparable to the sn ia heating rate of an entire galaxy ( see fig .
thus , it could be that the gas in the very center of the galaxies does not partake in the global sn ia wind , but periodically accretes into the central black hole .
environment can also play a significant role in the gas content of early - type galaxies .
xmm - newton and chandra observations of early - type galaxies in the coma cluster have shown that these galaxies can be either under luminous or over luminous in x - rays relative to field galaxies with comparable optical luminosities due to the competing effects of adiabatic compression , ram pressure stripping and thermal evaporation ( vikhlinin et al .
2001 ; finoguenov & miniati 2004 ; hornschemeier et al . 2005 ) . due to the large distance of the coma cluster ,
these observations were only sensitive to the x - ray properties of luminous early - type galaxies which have massive hydrostatic coronae in regions outside of dense groups or clusters .
low luminosity field galaxies , however , should have strong winds due to the lack of a confining ambient medium . ngc4552 is located in the densest environment of any of the galaxies in our sample and clearly shows the effects of its environment .
it is located @xmath22 ( a projected distance of 360 kpc ) from the center of the virgo cluster and has the highest gas content and hosts the most radio luminous agn in our sample .
the chandra image shows that ngc 4552 has a leading cold front and a trailing wake of ram pressure stripped gas ( machacek et al . 2005 ) .
based on the pressure jump across the cold front , machacek et al .
estimated that ngc4552 is traveling at 1600 km s@xmath12 .
the average gas density and pressure in ngc4552 derived from the values in table 5 are @xmath23 @xmath24 and @xmath25 kev @xmath24 . using the results in machacek et al . , the ram pressure of the virgo gas at the leading cold front in ngc4552 is @xmath26 kev @xmath24 , which is greater than the average thermal pressure of the gas , but less than the central pressure of @xmath27 kev @xmath24 .
these calculations are in good agreement with the fact that some of the outer gas in ngc452 is being stripped as it falls into the virgo cluster , but not the central gas .
based on the inferred velocity of ngc4552 , it should have traveled approximately 2 mpc during the time it has taken to accumulate its present gas mass .
the rosat all - sky survey data show that the gas in virgo follows a @xmath28 profile ( schindler et al 1999 ) . assuming the virgo gas is isothermal and a radial trajectory for ngc4552
, the thermal pressure of the virgo gas surrounding ngc4552 would have increased by a factor of 20 during this time and the ram pressure by an even greater factor due to the increasing infall velocity .
thus , the ambient pressure during most of this time was sufficient to suppress the formation of a galactic wind , but insufficient to strip the gas from the galaxy .
this would have lead to a substantial increase in gas mass which could have triggered the agn outburst .
we have presented a systematic analysis of chandra observations of 18 low - luminosity early - type galaxies .
emission from lmxbs is very significant in these galaxies and comprises between 30 to 95% of the total 0.5 - 2.0 kev emission .
we find that the combined x - ray luminosity of lmxbs ( resolved plus the power - law component of the unresolved emission ) scales approximately linearly with @xmath29 among the galaxies in our sample with an average @xmath30 ratio comparable to that found in more luminous early - type galaxies . the gas temperature in these galaxies varies from 0.2 to 0.8 kev and roughly follows the observed trend between gas temperature and stellar velocity dispersion among more luminous early - type galaxies , albeit with significant scatter .
the ratio of energy per unit mass in the galaxies to that in the gas , @xmath31 , varies from 0.3 to 1.0 , indicating the presence of significant non - gravitational heating in most of the galaxies .
we do not find any correlation between gas temperature and @xmath15 , unlike more luminous galaxies .
all of the galaxies in our sample are gas poor with @xmath32 , compared to @xmath33 in more luminous early - type galaxies .
the observed gas mass is much less than that expected from the accumulation of stellar mass loss over the lifetime of the galaxies .
based on recent estimates of the stellar mass loss rate in early - type galaxies , the time required to accumulate the observed gas mass is @xmath34 yr . due to the low gas densities , the cooling time of the gas is significantly longer that the replenishment time .
thus , most of the gas shed by evolving stars in these galaxies could not have condensed out of the hot phase and formed stars , and must have been expelled from the galaxies in a wind .
the only exception to this is the most optically luminous galaxy in our sample , ngc4552 .
based on recent estimates of the sn ia rate and mechanical agn heating rate in early - type galaxies , we find that , on average , heating by type ia supernova should dominate over agn heating in galaxies with @xmath13 . using the stellar mass loss rate in athey et al .
( 2002 ) and the sn ia rate in cappellaro et al .
( 1999 ) , we find that sn ia are energetically capable of driving galactic winds in low - luminosity early - type galaxies even if the present sn ia rate has been overestimated by a factor of two or the stellar mass loss rate has been underestimated by a factor of two .
we also find that most of the supernova energy must be consumed in lifting the gas out of the gravitational potential well of the galaxies , with little energy converted into thermal or bulk kinetic energy of the gas .
if heating by sn ia were the dominant heating mechanism at all times , it would be difficult to account for the large scatter in gas properties among the galaxies in our sample . even though our calculations show that the time - averaged agn heating rate is less than the supernova heating rate , the episodic nature of agn outbursts could produce periods when agn heating dominates .
one possible explanation for the large scatter in gas properties is that galaxies with very little gas at the present time recently experienced a significant agn outburst that produced a rapid expulsion of the gas .
six of the galaxies in our sample were detected in the nvss ( condon et al .
1998 ) , however , we do not find any correlations between @xmath14 and radio power .
the lack of a correlation may simply result from the time delay between when the agn reaches its peak radio power and when the gas is expelled from the galaxy .
in fact , the most radio luminous agn in our sample has the largest gas mass .
another possible source for the observed scatter in gas properties among early - type galaxies is variations in the ambient gas pressure surrounding the galaxies which would impact their ability to develop galactic winds .
we would like to thank the referee , mike loewenstein , for his many useful comments .
we also had many useful discussions of our work with ewan osullivan and jan vrtilek .
this work was supported by nasa grant ar4 - 5016x .
angelini , l. et .
2001 , apj , 557 , l35 .
athey , a. , bregman , j. , bregman , j. , temi , p. & sauvage , m. 2002 , apj , 571 , 272 .
bell , e. & de jong , r. 2001 , apj , 550 , 212 .
best , p. , kaiser , c. , heckman , t. & kauffmann , g. 2006 ( astro - ph 0602171 ) .
birzan , l. , rafferty , d. , mcnamara , b. , wise , m. & nulsen , p. 2004 , apj , 607 , 800 .
blanton , e. , sarazin , c. & irwin , j. 2001 , apj , 552 , 106 .
blanton , e. , sarazin , c. & mcnamara 2003 , apj , 585 , 227 .
bower , r. , benson , a. , malbon , r. , helly , j. , frenk , c. , baugh , c. , cole , s. & lacey , c. 2005 , ( astro - ph 0511338 ) .
brighenti , f. & mathews , w. 1999 , apj , 512 , 65 .
cappellaro , e. , evans , r. & turatto , m. 1999 , a&a , 351 , 459 .
ciotti , l. , deercole , a. , pellegrini , s. & renzini , a. 1991 , apj , 376 , 380 .
colbert , e. , heckman , t. , ptak , a. , strickland , d. weaver , k. 2004 , apj , 602 , 231 .
condon , j. cotton , w. , greisen , e. , yin , q. , perley , r. , taylor , g. & broderick , j. 1998 , aj , 115 , 1693 .
croton , d. , springel , v. , white , s. , delucia , g. , frenk , c. , gao , l. , jenkins , a. , kauffmann , g. , navarro , j. & yoshida , n. 2006 , mnras , 365 , 11 .
david , l. , forman , w. & jones , c. 1990 , apj , 359 , 29 .
david , l. , forman , w. & jones , c. 1991 , apj , 369 , 121 . de vaucouleurs , g. , de vaucouleurs a. , corwin , h. , buta , r. , paturel , g. & fouque , p. 1991
, third reference catalog of bright galaxies ( springer - verlag : berlin ) . di stefano , r. & kong , a. 2004 , apj , 609 , 710 .
faber , s. & gallagher , j. 1976 , apj , 204 , 365 , faber , s. & jackson , r. 1976 , apj , 204 , 668 .
faber , s. , wegner , g. , burstein , d. , davis , r. , dressler , a. , lynden - bell , d. & terlevich , r. 1989 , apj suppl , 69 , 763 .
fabian , a. , sanders , j , allen , s. , crawford , c. , iwasawa , k. , johnstone , r. , schmidt , w. & taylor g. 2003 , mnras , 344 , 43 .
finoguenov , a. & jones , c. 2002 , apj , 574 , 754 .
finoguenov , a. & miniati , r. 2004 , a&a , 418 , l21 .
forman , w. , jones , c. & tucker , w. 1985 , apj , 293 , 102 .
forman , w. , nulsen , p. , heinz , s. , owen , f. , eilek , j , , vikhlinin , a. , markevitch , m. , kraft , r. , churazov , e. , jones c. 2005 , apj , 635 , 894 .
fukazawa , y. , botoya - nonesa , j. , pu , j. , ohto , a. & kawano , n. 2006 , apj , 636 , 698 .
gebhardt , k. et al .
2000 , apj , 539 , 13 .
gilfanov , m. 2004 , mnras , 349 , 146 .
grevesse , n. & sauval , a.j .
1998 , space science rev . , 85 , 161 .
hornschemeier , a , mobasher , b. , alexander , d. , bauer , f. , bautz , m. , hammer , d. , poggianti , b. 2005 , ( astro - ph 0512547 ) .
humphrey , p. , buote , d. , gastaldello , f. , zappacosta , l. , bullock , j. , brighenti , f. & mathews , w. 2006 ( astro - ph 0601301 ) .
irwin , j. , athey , e. & bregman , j. 2003 , 587 , 356 .
jarrett , t. , chester , t. , cutri , r. , schneider , s. & huchra , j. 2003 , aj , 125 , 525 .
jones , c. , forman , w. , vikhlinin , a. , markevitch , m. , david , l. , warmflash , a. , murray , s. & nulsen , p. 2002
, apj , 567 , 115 .
jones , c. et al .
2006 , ( in preparation ) .
jordan , a. et al .
2004 , apj , 613 , 279 .
kahabka , p. & van den heuvel , e. 1997 , ara&a , 35 , 69 .
kim , d. , fabbiano , p. & trinchieri , g. 1992 , apj , 393 , 134 .
kim , d. & fabbiano , g. 2004 , apj , 611 , 846 .
kraft , r. , kregenow , j. , forman , w. , jones , c. , murray , s. 2001 , apj , 560 , 675 .
kraft , r. , vzquez , s. , forman , w. , jones , c. , murray , s. , hardcastle , m. , worrall , d. & churazov , e. 2003 , 129 , 146 .
lauer , t. et al .
2005 , aj , 129 , 2138 .
loewenstein , m. & mathews , w. 1987 , apj , 319 , 614 .
machacek , m. , jones , c. , forman , w. & nulsen , p. 2005
( astro - ph 0508588 ) .
mathews , w. & baker , j. , 1971 , apj , 170 , 241 .
mclure , r. & dunlop , j. 2002 , mnras , 331 , 705 .
mcnamara , b. et al .
2000 , apj , 534 , 135 .
navarro , j. , frenk , c. & white , s. 1996 , apj , 462 , 563 .
nulsen , p. , mcnamara , b. , wise , m. & david , l. 2005a , apj , 628 , 629 .
nulsen , p. , hambrick , d. , mcnamara , b. , rafferty , d. , birzan , l. , wise , m. , david , l. 2005b , apj , 625 , 9 .
osullivan , e. , forbes , d. & ponman t. 2001 , mnras , 328 , 461 .
osullivan , e. , ponman , t. & collins , r. 2003 , mnras , 340 , 1375 .
pellegrini , s. & fabbiano , g. 1994 , 429 , 105 .
pellegrini , s. 2005 , mnras , 364 , 169 .
pence , w. , snowden , s. , mukai , k. & kuntz , k. 2001 , apj , 561 , 189 .
rosati , p. et al .
2002 , apj , 566 , 667 .
sarazin , c. , irwin , j. & bregman , j. 2000 , apj , 544 , l101 .
sarazin , c. , irwin , j. & bregman , j. 2001 , apj , 556 . 533 .
sarazin , c. , kundu , a. , irwin , j. , sivakoff , g. , blanton , e. , randall , s. 2003 , apj , 595 , 743 .
schindler , s. , binggeli , b. & bohringer , h. 1999 , a & a , 343 , 420 .
swartz , d. , ghosh , k. , suleimanov , v. , tennant , a. & wu , k. 2002 , apj , 574 , 382 .
tammann , g. 1982 , in `` supernovae : a survey of current research ; proceedings of the advanced study institute '' , ( cambridge : england ) , p. 371 .
tonry , j. 2001 , apj , 546 , 681 .
trinchieri , g. & fabbiano , g. 1985 , apj , 296 , 447 .
vikhlinin , a. , markevitch , m. , forman , w. & jones , c. 2001 , apj , 555 , l87 .
wargelin , b. , markevitch , m. , juda , m. , kharchenko , v. , edgar & dalgarno , a. 2004 , apj , 607 , 596 .
white , s. & frenk , c. 1991 , apj , 379 , 52 .
white , r. , sarazin , c. & kulkarni , s. 2002 , apj , 571 , 23 .
wilson , c. , patel , s. , kouveliotou , c. , jonker , p. , van der klis , m. , lewin , w. , belloni , t. & mendez , m. 2003 , apj , 596 , 1220 .
consider a steady , spherical wind driven by energy injected in association with stellar mass loss . following mathews & baker ( 1971 )
, we assume that stellar mass loss is distributed smoothly and mixes well with the hot ism , so that we may treat the wind as a single , smooth fluid , with distributed mass injection .
the equation of mass conservation is then @xmath35 where @xmath36 is the radius , @xmath37 is the gas density , @xmath38 is its radial velocity and @xmath39 is the mass injection rate per unit volume due to stellar mass loss . assuming that the wind extends right to the center of the galaxy , this equation can be integrated to give the mass flow rate through a sphere of radius @xmath36 , @xmath40 where we have also defined @xmath41 , the mean stellar mass loss rate per unit volume within @xmath36 .
the momentum equation for the steady wind is @xmath42 where @xmath43 is the gas pressure and @xmath44 is the acceleration due to gravity ( positive inward ) . the energy equation may be written @xmath45 where @xmath46 is the specific enthalpy , @xmath47 , \label{eqn : enthalpy}\ ] ] @xmath48 is the ratio of specific heats , @xmath49 is the gravitational potential and @xmath50 is the total energy per unit mass that is deposited in the flow along with the stellar mass loss .
this is due principally to supernova heating , but includes kinetic energy due to random stellar motions , energy injected by planetary nebulae , etc .
following the usual practice , we assume that @xmath50 does not vary throughout the galaxy . integrating the energy equation gives @xmath51 4 \pi r'^2 \ , dr ' = { \dot m}w , \label{eqn : wdef}\ ] ] which also defines @xmath52 , a known function of the radius , and we have @xmath53 we ignore the effects of radiative energy loss , which would clearly increase the energy required to drive the wind .
these are generally small for the systems considered here . from equation ( [ eqn : mass ] ) , the density can be given in terms of the velocity as @xmath54 and from equations ( [ eqn : bernoulli ] ) and ( [ eqn : enthalpy ] ) we get @xmath55 these can be used to eliminate the pressure derivative from the momentum equation ( [ eqn : mom ] ) , giving , after some algebra , @xmath56 - \alpha \left [ { \gamma + 1\over 2 } v^2 + ( \gamma - 1 ) \epsilon \right ] , \label{eqn : spmom}\ ] ] where the sound speed , @xmath57 , is given by @xmath58 and the kepler speed , @xmath59 , is defined by @xmath60 .
note that in order to obtain this result , we have used @xmath61 and differentiated the definitions of @xmath62 ( equation [ eqn : mass ] ) and @xmath63 ( equation [ eqn : wdef ] ) .
the wind solutions starts with @xmath64 at @xmath65 , passes through a sonic point , where @xmath66 , and ends up as a freely expanding , supersonic flow for @xmath67 . at the sonic point
@xmath68 ( equations [ eqn : temp ] and [ eqn : sound ] ) , so that @xmath69 , \label{eqn : atsp}\ ] ] where @xmath70 is the radius of the sonic point .
in order for the solution to pass through the sonic point , the right hand side of equation ( [ eqn : spmom ] ) must vanish there . after using equation ( [ eqn : atsp ] ) to eliminate the velocity ,
the resulting condition may be written @xmath71 [ w({r_{\rm s } } ) - \phi({r_{\rm s } } ) ] = \alpha({r_{\rm s } } ) \epsilon + { { \overline{\alpha}}{v_{\rm k}}^2 \over 3 ( \gamma - 1)}. \label{eqn : sola}\ ] ] the terms on the right of this equation are both positive and , since @xmath72 , equation ( [ eqn : bernoulli ] ) requires @xmath73 too .
thus @xmath74 defining @xmath75 from the definition of @xmath63 ( equation [ eqn : wdef ] ) , we have @xmath76 .
using this in equation ( [ eqn : sola ] ) and solving for @xmath50 gives @xmath77 = \left [ { 4 { \overline{\alpha}}({r_{\rm s } } ) \over 3 ( \gamma + 1 ) } - \alpha({r_{\rm s } } ) \right ] [ \phi({r_{\rm s } } ) - { \overline{\phi}}({r_{\rm s } } ) ] + { { \overline{\alpha}}({r_{\rm s } } ) \over 3 ( \gamma - 1 ) } { v_{\rm k}}^2({r_{\rm s}}).\ ] ] since @xmath49 is an increasing function of the radius , @xmath78 , and we showed that the factor multiplying it is positive above ( [ eqn : ineq ] ) .
thus , we must also have @xmath79 under the usual assumption that @xmath80 is proportional to the stellar density , this condition requires the mean density to be substantially greater than the local density at the sonic point .
thus , the sonic point must occur well outside the dense stellar core of a typical galaxy .
we can now rewrite the result for @xmath50 as @xmath81_{r = { r_{\rm s } } } , \label{eqn : result}\ ] ] knowing that all of the factors in the term multiplied by @xmath80 are positive . for a galaxy model with specified potential and stellar mass loss rate , @xmath80
, this expression enables us to determine the specific energy injection rate , @xmath50 , from the location of the sonic point . alternatively ,
if the energy injection rate is known , this equation may be solved to find the location of the sonic point .
in particular , in the limit of large energy input , the sonic point approaches the radius that makes the denominator of the second term vanish , i.e. , that makes the inequality ( [ eqn : large ] ) an indentity .
for the remainder of this discussion , we assume that the stellar mass loss rate is proportional to the stellar density . inside a galaxy ,
the right hand side of the expression for @xmath50 is generally a decreasing function of the sonic radius , @xmath70 , so that the energy input required to drive a wind is minimized when the sonic point occurs at large @xmath36 .
if the stellar density falls rapidly to zero at the edge of the galaxy , @xmath80 falls to zero quickly , while the other terms in this expression are continuous .
thus there is a rapid reduction in the energy required to drive the wind as the sonic point moves across the outer edge of the galaxy . as a result ,
the sonic point occurs close to the edge of the stellar distribution for a range of values of @xmath50 . beyond the edge of the stars
@xmath82 so that @xmath83 is constant ( from the definition [ eqn : pbdef ] )
. the remaining variable terms in ( [ eqn : result ] ) , @xmath84 $ ] , may increase or decrease with radius , depending on the value of @xmath48 and the distribution of the gravitating matter ( which may extend beyond the stars ) . for @xmath85 ,
these terms reduce to @xmath86 , which is a non - decreasing function of @xmath36 ( since @xmath87 , where @xmath88 is the gravitating mass within @xmath36 ) , so that the global minimum of the expression ( [ eqn : result ] ) occurs at a value of @xmath70 close to , or immediately outside the edge of the stellar distribution .
however , the local condition for the sonic point ( [ eqn : result ] ) does not guarantee that the wind solution can be extended to infinity . beyond the edge of the stars , @xmath52 is also constant , and it is evident from equation ( [ eqn : bernoulli ] ) that it necessary that @xmath89 for the wind solution to extend to @xmath90 , an essential requirement for a steady wind .
thus the global minimum value of @xmath50 required for a wind when @xmath91 is @xmath92 similar considerations apply for @xmath93 , except that the energy requirement may now be minimized for @xmath94 in equation ( [ eqn : result ] ) .
thus we arrive at the same global value for @xmath95 .
note that , for @xmath91 , the wind solution with @xmath96 attains a constant mach number beyond the edge of the gravitating matter , rather than having the mach number approach infinity .
for @xmath93 , when the energy requirement is minimized for @xmath97 , the flow would be subsonic everywhere . in that case , the effect of a disturbance at finite @xmath36 is felt throughout the wind flow . in either case , the wind solution is fragile .
this is not surprising for a wind driven with the absolute minimum energy .
it emphasizes that the minimum energy requirement is just that . in reality
, greater energy is required to sustain a wind . to evaluate @xmath98
, we assume that the gravitational potential has the nfw form , @xmath99 , \qquad\qquad { \rm for\ } r < ac,\ ] ] where @xmath100 is the scale length , @xmath101 is the scale density and @xmath102 is the concentration parameter .
the stars , hence the mass injection rate , are assumed to have the king model distribution @xmath103 where @xmath104 is the core radius . using the definition ( [ eqn : pbdef ] ) ,
we then get @xmath105 , \\\end{aligned}\ ] ] where @xmath106 , @xmath107 and the stellar distribution is cut off at @xmath108 .
the normalizing factor for the nfw potential , @xmath109 , was determined from the assumption that the maximum rotation speed is related to the line - of - sight velocity dispersion by @xmath110 , giving @xmath111 with @xmath112 . equating the mean density of the halo to @xmath113 times
the critical density then gives the scale length , @xmath114 \over \chi c^3},\ ] ] where @xmath115 is the hubble constant .
parameters for the king model are set to match photometric properties of the galaxies . | we have analyzed chandra observations of 18 low - luminosity early - type galaxies with @xmath0 .
thermal emission from hot gas with temperatures between 0.2 and 0.8 kev comprises 5 - 70% of the total 0.5 - 2.0 kev emission from these galaxies .
we find that the total x - ray luminosity from lmxbs ( resolved plus the power - law component of the unresolved emission ) scales roughly linearly with the k - band luminosity of the galaxies with a normalization comparable to that found in more luminous early - type galaxies .
all of the galaxies in our sample are gas poor with gas masses much less than that expected from the accumulation of stellar mass loss over the life time of the galaxies .
the average ratio of gas mass to stellar mass in our sample is @xmath1 , compared to more luminous early - type galaxies which typically have @xmath2 .
the time required to accumulate the observed gas mass from stellar mass loss in these galaxies is typically @xmath3 yr . since the cooling time of the gas is longer than the replenishment time , the gas can not be condensing out of the hot phase and forming stars , implying that the gas is most likely being expelled from these galaxies in a wind . the one exception to this is ngc4552 , which is the most optically luminous galaxy in our sample and has the highest gas content . using recent estimates of the type ia supernova rate and agn heating rate in early - type galaxies , we find that , on average , heating by type ia supernovae should exceed agn heating in galaxies with @xmath0
. we also find that heating by type ia supernovae is energetically sufficient to drive winds in these galaxies , even if the present type ia supernova rate is overestimated by a factor of two or the present stellar mass loss rate is underestimated by a factor of two .
there is significant scatter in the gas properties of galaxies with comparable optical luminosities .
nearly continuous heating by type ia supernovae alone can not account for the large scatter in observed gas mass .
we suggest that the scatter in gas properties could arise from periodic agn outbursts , during which time agn heating dominates over type ia supernova heating , or environmental factors , such as a high pressure environment which would suppress the formation of strong galactic winds . |
the investigation of high - energy electromagnetic processes in crystals is of interest not only from the viewpoint of underlying physics but also from the viewpoint of practical applications . from the point of view of controlling the parameters of various processes in a medium , it is of interest to investigate the influence of external fields , such as acoustic waves , temperature gradient etc .
, on the corresponding characteristics .
the considerations of concrete processes , such as diffraction radiation mkrtchlsh , transition radiation @xcite , parametric x - radiation @xcite , channelling radiation @xcite , bremsstrahlung by high - energy electrons @xcite , have shown that the external fields can essentially change the angular - frequency characteristics of the radiation intensities .
recently there has been broad interest to compact crystalline undulators with periodically deformed crystallographic planes an efficient source of high energy photons @xcite ( for a review with more complete list of references see @xcite ) .
the coherent bremsstrahlung of high - energy electrons moving in a crystal is one of the most effective methods to produce intense beams of highly polarized and monochromatic photons .
such radiation has a number of remarkable properties and at present it has found many important applications .
this motivates the importance of investigations for various mechanisms of controlling the radiation parameters . in @xcite
we have discussed the influence of hypersonic waves excited in a crystal on the process of the bremsstrahlung of high - energy electrons . to have an essential influence of the acoustic wave high - frequency hypersound is needed .
usually this type of waves is excited by high - frequency electromagnetic field through the piezoelectric effect in crystals with a complex base . in the paper @xcite , we have generalized the results of @xcite for crystals with a complex base and for acoustic waves with an arbitrary profile . for the experimental detection of final particles in the process of coherent bremsstrahlung
it is important to know their angular distribution . in the present paper the angular distribution of photons in the coherent bremsstrahlung in crystals
is investigated in the presence of hypersonic wave .
the numerical calculations are carried out for the quartz single crystal and for the electrons of energy 20 gev .
the paper is organized as follows . in the next section
we derive the general formula for the coherent part of the bremsstrahlung cross - section averaged on thermal fluctuations and the conditions are specified under which the influence of the deformation field can be considerable . in sec .
3 the analysis of the general formula is presanted in the cases when the electron enters into the crystal at small angles with respect to crystallographic axes or planes and the results of the numerical calculations for the cross - section as a function of the angle between the momenta of electron and photon .
4 summarizes the main results of the paper .
we consider the bremsstrahlung by high - energy electrons in a single crystal .
the corresponding cross section can be obtained from the cross section for the bremsstrahlung on a single atom derived in @xcite .
the formulae @xcite practically can not be used for the investigation of bremsstrahlung in case of potentials of a complex form , for example in a crystal , as errors of approximations , admissible in the case of a single scattering center , can become inadmissible greater for many centers . by making use of the results derived in @xcite ,
the differential cross section for the bremsstrahlung on a single atom is presented in the form ( the system of units @xmath1 is used)@xmath2 \frac{% \left\vert u\left ( \mathbf{q}\right ) \right\vert ^{2}}{\sqrt{1-y^{2 } } } \notag \\ & = & \left\vert u\left ( \mathbf{q}\right ) \right\vert ^{2}\sigma _ { 0}\left ( \mathbf{q,}y\right ) , \label{sigindatom}\end{aligned}\]]where @xmath3 is the electron charge , @xmath4 , @xmath5 , @xmath6 , are the energies of photon , initial and final electrons respectively , @xmath7 , @xmath8 is the mass of electron , @xmath9 and @xmath10 are the components of the vector of momentum transfer @xmath11 , @xmath12 ( @xmath13 , @xmath14 , @xmath15 are the momenta of photon , initial and final electrons respectively ) , parallel and perpendicular to the direction of the electron momentum , @xmath16 is the fourier transform of the atom potential .
the variable @xmath17 is expressed in terms of the angle @xmath18 between the momenta @xmath19 and @xmath13 by the following relation : @xmath20the regions of variables @xmath21 , @xmath22 , @xmath17 in cross - section ( [ sigindatom ] ) are as follows @xcite : @xmath23 the differential cross - section for the bremsstrahlung in a crystal can be written in the form
@xcite @xmath24 where @xmath25 is the position of an atom in the crystal . in the discussion that follows
, the collective index @xmath26 enumerates the elementary cell and the subscript @xmath27 enumerates the atoms in a given cell of a crystal . here
@xmath11 is the momentum transferred to the crystal , @xmath28 and the differential cross - section in a crystal given by ( [ sigkrysdef ] ) , differs from the cross - section on an isolated atom by the interference factor which is responsible for coherent effects arising due to the periodical arrangement of the atoms in the crystal . after averaging on thermal fluctuations ,
the cross - section is written in the form ( see , for instance @xcite for the case of a crystal with a simple cell)@xmath29where @xmath30 is the number of cells , @xmath31 determine the equilibrium positions of the atoms , @xmath32 is the temperature - dependent mean - squared amplitude of the thermal vibrations of the @xmath27th atom , @xmath33 is the corresponding debye - waller factor . in formula ( sigtermik )
the first term in figure braces does not depend on the direction of the vector @xmath13 and determines the contribution of incoherent effects .
the contribution of coherent effects is presented by the second term . by taking into account the formula ( [ sigindatom ] ) for the cross - section on a single atom , in the region @xmath34
the corresponding part of the cross - section in a crystal can be presented in the form @xmath35
\frac{1}{\sqrt{1-y^{2}}}\left\vert \sum_{n , j}u_{\mathbf{q}}^{\left ( j\right ) } e^{i\mathbf{qr}_{n0}^{\left ( j\right ) } } e^{-\frac{1}{2}q^{2}% \overline{u_{t}^{\left ( j\right ) 2}}}\right\vert ^{2 } \label{sigccrysfac}\ ] ] when external influences are present ( for example , in the form of acoustic waves ) the positions of atoms in the crystal can be written as @xmath36 , where @xmath37 determines the equilibrium position of an atom in the situation without deformation , @xmath38 is the displacement of the atom caused by the external influence .
we consider deformations with the periodical structure : @xmath39where @xmath40 and @xmath41 are the amplitude and wave vector corresponding to the deformation field , @xmath42 is an arbitrary function with the period @xmath43 , @xmath44 @xmath45 . in the discussion that folows , we assume that @xmath46 . further following @xcite and
having made similar steps we receive the following formula for the differential cross section of coherent bremsstrahlung:@xmath47 \frac{\left\vert f_{m}\left ( \mathbf{g}_{m}\mathbf{u}_{0}\right ) \right\vert ^{2}\left\vert s\left ( \mathbf{g}_{m},\mathbf{g}\right ) \right\vert ^{2}}{\sqrt{1-y^{2 } } } , \label{sigcohgeneral}\]]where @xmath48 is the number of atoms in the crystal @xmath49 , @xmath50 is the reciprocal lattice vector , @xmath51 stands for the momentum transfer to the external field , @xmath52 and @xmath53 are the components of the vector @xmath54 .
the function @xmath55 is defined by relation@xmath56 is the factor determined by the structure of the elementary cell . for a lattice with a complex cell
the coordinates of the atoms are written as @xmath57 , with @xmath58 being the positions of the atoms for one of primitive lattices , and @xmath59 are the equilibrium positions for other atoms inside @xmath26-th elementary cell with respect to @xmath60 .
now the relation between the variables @xmath17 and @xmath18 is written in the form:@xmath61 ^{\frac{1}{2 } } } \label{ydefincrys}\ ] ] the regions of variables in cross - section ( [ sigkrysdef ] ) are@xmath62for sinusoidal deformation field , @xmath63 , one has the fourier - transform@xmath64with the bessel function @xmath65 .
the formula for the pair creation in an undeformed crystal is obtained from ( [ sigcohgeneral ] ) taking @xmath66 . in this limit , the contribution of the term with @xmath67 remains only with @xmath68 .
now we see that formula ( [ sigcohgeneral ] ) differs from the formula in an undeformed crystal by the replacement @xmath69 , and by the additional summation over @xmath70 with the weights @xmath71 .
this corresponds to the presence of an additional one - dimensional superlattice with the period @xmath72 and the reciprocal lattice vector @xmath73 , @xmath74 .
as the main contribution into the cross - section comes from the terms with @xmath75 , the influence of the deformation field may be considerable if @xmath76 . combining this with the previous estimates , we find the following condition : @xmath77 . at high energies
one has @xmath78 and this condition can be consistent with the condition @xmath79 . in the presence of the deformation field the number of possibilities to satisfy the condition @xmath80 in the summation of formula ( [ sigcohgeneral ] ) increases due to the term @xmath81 in the expression for @xmath82 .
this leads to the appearance of additional peaks in the angular distribution of the radiated positrons .
after the integration of ( [ sigcohgeneral ] ) over @xmath17 , due to these additional peaks , there can be an enhancement of the cross - section of the process @xcite .
now , we consider the most interesting case when the electron enters into the crystal at small angle @xmath83 with respect to the crystallographic @xmath84-axis of the orthogonal lattice .
the corresponding reciprocal lattice vector components are @xmath85 @xmath86 , where @xmath87 @xmath88 are the lattice constants in the corresponding directions . for the longitudinal component we can write @xmath89where
@xmath90 is the angle between the projection of the vector @xmath91 on the plane @xmath92 and axis @xmath17 . for small angles
@xmath83 the main contribution into the cross - section comes from the summands with @xmath93 . having made the replacement of variable @xmath94 using the formula ( [ ydefincrys ] ) from formula ( sigcohgeneral ) one finds@xmath95 \notag \\ & & \times \frac{\left\vert f_{m}\left ( \mathbf{g}_{m}\mathbf{u}_{0}\right ) \right\vert ^{2}\left\vert s\left ( \mathbf{g}_{m},\mathbf{g}\right ) \right\vert ^{2}}{\sqrt{1-y^{2}\left ( \theta _ { \gamma } \right ) } } \frac{% \epsilon _ { 1}\theta _ { \gamma } /m_{e}}{\left ( g_{\perp } /m_{e}\right ) \left ( g_{m\parallel } /\delta
-1-g_{\perp } ^{2}/\left ( 2\epsilon _ { 1}\delta \right ) \right ) ^{\frac{1}{2 } } } , \label{sigcasegz0}\end{aligned}\]]where the notation @xmath96 is intrduced in accordance with:@xmath97
^{2}}{% g_{m\parallel } /\delta
-1-g_{m\perp } ^{2}/\left ( 2\epsilon _ { 1}\delta \right ) } .
\label{y2casegz0}\]]in ( [ sigcasegz0 ] ) @xmath98 , and the summation goes over the region @xmath99
@xmath100 with @xmath101note that in the argument of the functions @xmath102 and @xmath103 we have @xmath104 .
we now assume that the electron enters into the crystal at small angle @xmath105 with respect to the crystallographic axis @xmath84 and near the crystallographic plane @xmath106 ( the angle @xmath90 is small ) . in this case with the change of @xmath107 , the sum over @xmath108 and @xmath109 will drop sets of terms which lead to the abrupt change of the corresponding cross - section .
two cases have to distinguish . under the condition @xmath110 , in eq .
( [ sigcasegz0 ] ) for the longitudinal component , one has @xmath111 the formula ( [ sigcasegz0 ] )
can be further simplified under the assumption @xmath112 . in this case , in the argument of the function @xmath102 , one has @xmath113 and we obtain the formula @xmath114 \notag \\ & & \frac{\left\vert f_{m}\left ( g_{y}u_{y0}\right ) \right\vert ^{2}\left\vert s\left ( \mathbf{g}_{m},\mathbf{g}\right ) \right\vert ^{2}}{\sqrt{% 1-y^{2}\left ( \theta _ { \gamma } \right ) } } \frac{\varepsilon _ { 1}\theta _ { \gamma } /m_{e}}{\left ( g_{\perp } /m_{e}\right ) \left [ g_{m\parallel } /\delta
-1-g_{\perp } ^{2}/\left ( 2\epsilon _ { 1}\delta \right ) \right ] ^{% \frac{1}{2}}}. \label{sigsumgxgy26}\end{aligned}\ ] ] in the second case , we assume that @xmath115 . now the main contribution into the sum in eq .
( [ sigcasegz0 ] ) comes from terms with @xmath116 and summations remain over @xmath70 and @xmath117 , @xmath118 .
the formula for the cross - section takes the form@xmath119 \notag \\ & & \frac{\left\vert f_{m}\left ( \mathbf{g}_{m}\mathbf{u}_{0}\right ) \right\vert ^{2}\left\vert s\left ( \mathbf{g}_{m},\mathbf{g}\right ) \right\vert ^{2}}{\sqrt{1-y^{2}\left ( \theta _ { \gamma } \right ) } } \frac{% \epsilon _ { 1}\theta _ { \gamma } /m_{e}}{\left ( g_{m\perp } /m_{e}\right ) \left [ g_{m\parallel } /\delta
-1-g_{m\perp } ^{2}/\left ( 2\epsilon _ { 1}\delta \right ) \right ] ^{\frac{1}{2 } } } , \label{sigsumg127}\end{aligned}\]]where @xmath120and the summation goes over the values @xmath70 and @xmath117 satisfying the condition @xmath121 .
we have carried out numerical calculations for the coherent bremsstrahlung cross - section for various values of parameters in the case of @xmath122 single crystal at zero temperature . to deal with an orthogonal lattice , we choose as an elementary cell the cell including 6 atoms of silicon and 12 atoms of oxygen ( shrauf elementary cell @xcite ) . for
this choice the @xmath17 and @xmath84 axes of the orthogonal coordinate system @xmath123 coincide with the standard @xmath124 and @xmath125 -axes of the quartz crystal , whereas the angle between the axes @xmath126 and @xmath127 is equal to @xmath128 . for the potentials of atoms we take moliere parametrization
with@xmath129where @xmath130 @xmath131 @xmath132 is the screening radius for the @xmath27-th atom in the elementary cell . [
cols="^,^,^ " , ] c
the present paper is devoted to the investigation of the angular distribution of the electron in the bremsstrahlung process by high - energy electrons in a crystal with a complex lattice base in the presence of deformation field of an arbitrary periodic profile .
the latter can be induced , for example , by acoustic waves .
the influence of the deformation field can serve as a possible mechanism to control the angular - energetic characteristics of the created particles .
the importance of this is motivated by that the coherent bremsstrahlung of high - energy electrons moving in a crystal is one of the most effective methods to produce intense beams of highly polarized and monochromatic photons . in a crystal
the cross - section is a sum of coherent and incoherent parts .
the coherent part of the cross - section per single atom , averaged on thermal fluctuations , is given by formula ( [ sigcohgeneral ] ) . in this formula the factor @xmath133 is determined by the function describing the displacement of the atoms due to the deformation field , and the factor @xmath134 is determined by the structure of the crystal elementary cell .
compared with the cross - section in an undeformed crystal , formula ( [ sigcohgeneral ] ) contains an additional summation over the reciprocal lattice vector of the one - dimensional superlattice induced by the deformation field .
we have argued that the influence of the deformation field on the cross - section can be remarkable under the condition @xmath135 .
note that for the deformation with @xmath136 this condition is less restrictive than the naively expected one @xmath137 .
the role of coherence effects in the pair creation cross - section is essential when the photon enters into the crystal at small angles with respect to a crystallographic axis . in this case
the main contribution into the coherent part of the cross - section comes from the crystallographic planes , parallel to the chosen axis ( axis @xmath84 in our consideration ) .
the behavior of this cross - section as a function on the positron energy essentially depends on the angle @xmath83 between the projection of the electron momentum on the plane @xmath138 and @xmath17-axis .
when the electron enters into the crystal near a crystallographic plane , two cases have to be distinguished .
for the first one @xmath139 the formula ( [ sigcasegz0 ] ) is further simplified to the form ( [ sigsumgxgy26 ] ) under the assumption @xmath140 . in the second case one
has @xmath141 , and the main contribution into the cross - section comes from the crystallographic planes parallel to the incidence plane .
the corresponding formula for the cross - section takes the form ( [ sigsumg127 ] ) .
the numerical calculations for the cross - section are carried out in the case of sio@xmath142 single crystal with the moliere parametrization of the screened atomic potentials and for the deformation field generated by the transversal acoustic wave of @xmath103 - type with frequency 5 ghz .
examples of numerical results are depicted in figures .
the numerical calculations for values of the parameters in the problem when one has an enhancement of the cross - section show that the presence of the deformation field leads to the appearance of additional peaks in the angular distribution of the radiated photon even for such ranges of values of an angle of a photon , where due to the requirement @xmath143 the cross - section is zero when deformation is absent . this can be used to control the parameters of the photon sources .
i am grateful to aram saharian for valuable discussions and suggestions . 99 a. r. mkrtchyan , l. sh .
grigoryan , a. n. didenko , a. a. saharian , a. g. mkrtchyan , izv .
fizika * 24 * , 62 ( 1989 ) ; a. r. mkrtchyan , l. sh .
grigoryan , a. n. didenko , a. a. saharian , sov .
jtp * 61 * , 21 ( 1991 ) ; a.r .
mkrtchyan , l. sh .
grigoryan , a. a. saharian a. n. didenko , acustica * 75 * , 1984 ( 1991 ) ; a. a. saharian , a. r. mkrtchyan , l. v. gevorgian , l. sh .
grigoryan , b.v .
khachatryan , nucl .
instr . and meth .
* b173 * , 211 ( 2001 ) .
a. r. mkrtchyan , r. a. gasparyan , r. g. gabrielyan , phys . lett . *
a115 * , 410 ( 1986 ) ; jetp * 93 * 432 ( 1987 ) ; phys .
* a126 * 528 ( 1988 ) ; l. sh .
grigoryan et al . , rad .
eff . and def . in solids
* 152 * , 13 ( 2000 ) ; * 152 * , 225 ( 2000 ) ; * 152 * , 269 ( 2000 ) ; * 153 * , 221 ( 2001 ) ; * 153 * , 289 ( 2001 ) ; * 153 * , 307 ( 2001 ) ; l. sh .
grigoryan et al . , nucl .
instr . and meth .
* b173 * , 13 ( 2001 ) ; * b173 * , 184 ( 2001 ) ; l. sh .
grigoryan , a. h. mkrtchyan , h. f. khachatryan , v. u. tonoyan , w. wagner , nucl . instr . and meth.*b201 * , 25 ( 2003 ) .
a. v. korol , a. v. solovyov , w. greiner , j. phys .
* 24 * l45 ( 1998 ) ; a. v. korol , a. v. solovyov , w. greiner , int . j. mod
e * 8 * , 49 ( 1999 ) ; r. o. avakian , l. a. gevorgian , k. a. ispirian , r. k. ispirian , jetp lett . * 68 * , 467 ( 1998 ) ; q .-
r . zhang , int . j. mod
e * 8 * 493 ( 1999 ) ; u. mikkelsen , e. uggerhj , nucl .
instr . and meth .
* b160 * 435 ( 2000 ) ; a. v. korol , a. v. solovyov , w. greiner , j. phys .
g : nucl . part
* 27 * 95 ( 2001 ) ; r. o. avakian ; l. a. gevorgian , k. a. ispirian , r. k. ispirian , nucl .
instr . and meth .
* b173 * 112 ( 2001 ) ; a. v. korol , w. krause , a. v. solovyov , w. greiner , nucl .
instr . and meth .
* a483 * , 455 ( 2002 ) ; r. o. avakian , k. t. avetyan , k. a. ispirian , e. g. melikyan ; nucl . instr . and meth .
* a508 * 496 ( 2003 ) ; s. bellucci , s. bini , v. m. biryukov , yu .
chesnokov _ et al _ phys .
lett . * 90 * , 034801 ( 2003 ) ; s. bellucci , mod .
b * 19 * 85 ( 2005 ) ; s. bellucci , v. a. maisheev , nucl .
instr . and meth .
* b252 * 339 ( 2006 ) .
h. bethe , w. heitler , proc .
soc . , * 146 * , 83 ( 1934 ) ; h. bethe , l. maximon , phys .
, * 93 * , 768 ( 1954 ) ; h. davis , h. bethe , l. maximon , phys . rev . , * 93 * , 788 ( 1954 ) ; h. olsen , l. maximon , h. wergeland , phys .
, * 106 * , 27 ( 1957 ) ; v. n. baier , v. m. katkov , v. s. fadin , radiation of relativistic electrons ( atomizdat , 1973 ) ( in russian ) . | we investigate the angular distribution of photons in the coherent bremsstrahlung process by high - energy electrons in a periodically deformed single crystal with a complex base . the formula for the corresponding differential cross - section
is derived for an arbitrary deformation field .
the case is considered in detail when the electron enters into the crystal at small angles with respect to a crystallographic axis .
the results of the numerical calculations are presented for @xmath0 single crystal and moliere parameterization of the screened atomic potentials in the case of the deformation field generated by the acoustic wave of s -type .
_ keywords : _ interaction of particles with matter ; coherent bremsstrahlung ; physical effects of ultrasonics .
pacs nos . : 41.60.-m , 78.90.+t , 43.35.+d , 12.20.ds |
the sld resolution used in prolog may not be complete or efficient for programs in the presence of recursion .
for example , for a recursive definition of the transitive closure of a relation , a query may never terminate under sld resolution if the program contains left - recursion or the graph represented by the relation contains cycles even if no rule is left - recursive . for a natural definition of the fibonacci function
, the evaluation of a subgoal under sld resolution spawns an exponential number of subgoals , many of which are variants .
the lack of completeness and efficiency in evaluating recursive programs is problematic : novice programmers may lose confidence in writing declarative programs that terminate and real programmers have to reformulate a natural and declarative formulation to avoid these problems , resulting in cluttered programs .
tabling @xcite is a technique that can get rid of infinite loops for bounded - term - size programs and redundant computations in the execution of recursive programs .
the main idea of tabling is to memorize the answers to subgoals and use the answers to resolve their variant descendents .
tabling helps narrow the gap between declarative and procedural readings of logic programs .
it not only is useful in the problem domains that motivated its birth , such as program analysis @xcite , parsing @xcite , deductive databases @xcite , and theorem proving @xcite , but also has been found essential in several other problem domains such as model checking @xcite and logic - based probabilistic learning@xcite .
this idea of caching previously calculated solutions , called _ memoization _ , was first used to speed up the evaluation of functions @xcite .
oldt @xcite is the first resolution mechanism that accommodates the idea of tabling in logic programming and xsb is the first prolog system that successfully supports tabling @xcite .
tabling has become a practical technique thanks to the availability of large amounts of memory in computers .
it has become an embedded feature in a number of other logic programming systems such as b - prolog @xcite , mercury @xcite , tals @xcite , and yap @xcite .
oldt , and slg @xcite alike , is non - linear in the sense that the state of a consumer must be preserved before execution backtracks to its producer .
this non - linearity requires freezing stack segments @xcite or copying stack segments into a different area @xcite before backtracking takes place .
linear tabling is an alternative tabling scheme @xcite .
the main idea of linear tabling is to use iterative computation of looping subgoals rather than suspension and resumption of them as is done in oldt to compute fixpoints .
this basic idea dates back to the et * algorithm @xcite .
the dra method proposed in @xcite is based on the same idea but employs different strategies for handling looping subgoals and clauses . in linear tabling , a cluster of inter - dependent subgoals as represented by a _ top - most looping subgoal _
is iteratively evaluated until no subgoal in it can produce any new answers .
linear tabling is relatively easy to implement on top of a stack machine thanks to its linearity , and is more space efficient than oldt since the states of subgoals need not be preserved .
linear tabling is a framework from which different methods can be derived based on the strategies used in handling looping subgoals .
one decision concerns when answers are consumed and returned .
the _ lazy _ strategy postpones the consumption of answers until no answers can be produced .
it is in general space efficient because of its locality and is well suited for all - solution search programs . the _ eager _
strategy , in contrast , prefers answer consumption and return over production .
it is well suited for programs with cuts .
these two strategies have been compared in slg - wam as two scheduling strategies called _ local _ and _ single - stack _ @xcite .
this paper gives a comprehensive analysis of these two strategies and compares their performance experimentally .
linear tabling relies on iterative evaluation of top - most looping subgoals to compute fixpoints .
naive re - evaluation of all looping subgoals may be computationally expensive .
_ semi - naive optimization _ is an effective technique used in bottom - up evaluation of datalog programs @xcite .
it avoids redundant joins by ensuring that the join of the subgoals in the body of each rule must involve at least one new answer produced in the previous round .
the impact of semi - naive optimization on top - down evaluation had been unknown before @xcite . in this paper
, we also propose to introduce semi - naive optimization into linear tabling .
we have made efforts to properly tailor semi - naive optimization to linear tabling . in our semi - naive optimization ,
answers for each tabled subgoal are divided into three regions as in bottom - up evaluation , but answers are consumed sequentially until exhaustion not incrementally as in bottom - up evaluation so that answers produced in a round are consumed in the same round .
we have found that incremental consumption of answers does not fit linear tabling since it may require more iterations to reach fixpoints .
moreover , consuming answers incrementally may cause redundant consumption of answers .
we further propose a technique called _ early promotion _ of answers to reduce redundant consumption of answers .
our benchmarking shows that this technique gives significant speed - ups to some programs .
an efficient tabling system has been implemented in b - prolog , in which the lazy strategy is employed by default but the eager strategy can be used through declarations for subgoals that are in the scopes of cuts or are not required to return all the answers .
our tabling system not only consumes considerably less stack space than xsb for some programs but also compares favorably well in speed with xsb .
the theoretical framework of linear tabling is given in @xcite .
the main objective of this paper is to propose evaluation strategies and their optimizations for linear tabling .
the remainder of the paper is structured as follows : in the next section we define the terms used in this paper . in section 3
we give the linear tabling framework and the two answer consumption strategies . in section 4
we introduce semi - naive optimization into linear tabling and prove its completeness . in section 5
we describe the implementation of our tabling system and also show how to implement semi - naive optimization . in section 6
we compare the tabling strategies experimentally , evaluate the effectiveness of semi - naive optimization , and also compare the performance of b - prolog with xsb . in section 7
we survey the related work and in section 8 we conclude the paper .
in this section we give the definitions of the terms to make this paper as much self - contained as possible .
the reader is referred to @xcite for a description of sld resolution . in this paper
, we always assume the top - down strategy for selecting clauses and the left - to - right computation rule .
let @xmath0 be a program .
tabled predicates in @xmath0 are explicitly declared and all the other predicates are assumed to be non - tabled .
a subgoal of a tabled predicate is called a _ tabled subgoal_. tabled predicates are transformed into a form that facilitates execution : each rule ends with a dummy subgoal named @xmath1 where @xmath2 is the head , and each tabled predicate contains a dummy ending rule whose body contains only one subgoal named _
check_completion(h)_. for example , given the definition of the transitive closure of a relation , .... : -table p/2 .
p(x , y):-p(x , z),e(z , y ) .
p(x , y):-e(x , y ) .
.... the transformed predicate is as follows : .... p(x , y):-p(x , z),e(z , y),memo(p(x , y ) ) .
p(x , y):-e(x , y),memo(p(x , y ) ) .
p(x , y):-check_completion(p(x , y ) ) . .... a table is used to record subgoals and their answers . for each subgoal and its variants
, there is an entry in the table that stores the state of the subgoal ( e.g. , complete or not ) and an answer table for holding the answers generated for the subgoal .
initially , the answer table is empty .
let @xmath3 and @xmath4 be two terms with no shared variables .
the term @xmath3 _ subsumes _ @xmath4 if there exists a substitution @xmath5 such that @xmath6=@xmath4 .
the two terms @xmath3 and @xmath4 are called _ variants _ if they subsume each other .
let @xmath7 be a goal .
the first subgoal @xmath8 is called the _ selected subgoal _ of the goal . @xmath9 is _ derived _ from @xmath10 by using a tabled _ answer _
@xmath11 if there exists a unifier @xmath5 such that @xmath12 and @xmath13 .
@xmath9 is _ derived _ from @xmath10 by using a rule `` @xmath14 '' if @xmath15 and @xmath16 .
@xmath8 is said to be the _ parent _ of @xmath17 , ... , and @xmath18 .
the relation _ ancestor _ is defined recursively from the parent relation .
a tabled subgoal that occurs first in the construction of an sld tree is called a _ pioneer _ , and all subsequent variants are called _ followers _ of the pioneer .
let @xmath19 be a given goal , and @xmath20 be a _ derivation _ where each goal is derived from the goal immediately preceding it .
let @xmath21 be a sub - sequence of the derivation where @xmath22 and @xmath23 .
the sub - sequence forms a _ loop _ if @xmath24 and @xmath25 are variants . the subgoals @xmath24 and @xmath25 are called _ looping subgoals_. in particular , @xmath24 is called the _ pioneer looping subgoal _ and @xmath25 is called the _ follower looping subgoal _ of the loop . notice that the pioneer and follower looping subgoals are not required to have the ancestor - descendent relationship , and thus a derivation that contains two variant subgoals may not be a _ real _ loop .
consider , for example , the goal `` @xmath26 '' where @xmath27 is defined by facts .
the derivation `` @xmath26 '' @xmath28 @xmath29 is treated as a loop although the selected subgoal @xmath29 in the second goal is not a descendant of @xmath30 .
a subgoal @xmath24 is said to be _ dependent _ on another subgoal @xmath25 if @xmath25 occurs in a derived goal from @xmath24 , i.e. , @xmath31 .
two subgoals are said to be _
inter - dependent _ if they are dependent on each other .
inter - dependent subgoals constitute a _ cluster _ , which is called a _ strongly connected component _
elsewhere @xcite .
a subgoal in a cluster is called the _ top - most _ subgoal of the cluster if none of its ancestors is included in the cluster .
unless a cluster contains only a single subgoal , its top - most subgoal must also be a looping subgoal .
for example , the subgoals at the nodes in the sld tree in figure [ fig : loops ] constitute a cluster and the subgoal p at node 1 is the top - most looping subgoal of the cluster .
linear tabling takes a transformed program and a goal , and tries to find a path in the sld tree that leads to an empty goal .
the primitive @xmath32 is executed when a tabled subgoal @xmath24 is encountered .
just as in sld resolution , linear tabling explores the sld tree in a depth - first fashion , taking special actions when _
table_start(a ) _ , _ memo(a ) _ , and _
check_completion(a ) _ are encountered .
backtracking is done in exactly the same way as in sld resolution . when the current path reaches a dead end , meaning that no action can be taken on the selected subgoal , execution backtracks to the latest previous goal in the path and continues with an alternative branch .
when execution backtracks to the top - most looping subgoal of a cluster , however , we can not fail the subgoal even after all the alternative clauses have been tried .
in general , the evaluation of a top - most looping subgoal must be iterated until its fixpoint is reached .
we call each iteration of a top - most looping subgoal a _ round_. various linear tabling methods can be devised based on the framework . a linear tabling method
comprises strategies used in the three primitives : _
table_start(a ) _ , _ memo(a ) _ , and _ check_completion(a)_. in linear tabling , a pioneer subgoal has two roles : one is to produce answers into the table and the other is to return answers to its parent through its variables .
different strategies can be used to produce and return answers .
the _ lazy strategy _ gives priority to answer production and the _ eager strategy _ prefers answer consumption over production . in the following
we define the three primitives in both strategies .
the lazy strategy postpones the consumption of answers until no answers can be produced . in concrete , for top - most looping subgoals no answer is returned until they are complete , and for other pioneer subgoals answers are consumed only after all the rules have been tried .
this primitive is executed when a tabled subgoal @xmath24 is encountered . the subgoal @xmath24 is registered into the table if it is not registered yet . if @xmath24 s state is _ complete _ meaning that @xmath24 has been completely evaluated before , then @xmath24 is resolved by using the answers in the table . if @xmath24 is a pioneer , meaning that it is encountered for the first time in the current path , then different actions are taken depending on @xmath24 s state .
if @xmath24 s state is _ evaluated _ meaning that @xmath24 has occurred before in a different path during the current round , then it is resolved by using answers .
otherwise , if @xmath24 has never occurred before during the current round , it is resolved by using rules . in this way
, a pioneer subgoal needs to be evaluated only once in each round .
if @xmath24 is a follower of some ancestor @xmath33 , meaning that a loop has been encountered , must be an ancestor of @xmath24 under the lazy strategy .
] then it is resolved by using the answers in the table .
after the answers are exhausted , @xmath24 fails . failing @xmath24 is unsafe in general since it may have not returned all of its possible answers .
for this reason , the top - most looping subgoal of the cluster of @xmath24 needs be iterated until no new answer can be produced .
this primitive is executed when an answer is found for the tabled subgoal @xmath24 .
if the answer @xmath24 is already in the table , then just fail ; otherwise fail after the answer is added into the table .
the failure of _ memo _ postpones the return of answers until all rules have been tried .
this primitive is executed when the subgoal @xmath24 is being resolved by using rules and the dummy ending rule is being tried .
if @xmath24 has never occurred in a loop , then @xmath24 s state is set to _ complete _ and @xmath24 is failed after all the answers are consumed . if @xmath24 is a top - most looping subgoal , we check if any new answers are produced during the last iteration of the cluster under @xmath24 .
if so , @xmath24 is re - evaluated by calling @xmath32 after all the dependent subgoals s states are initialized .
otherwise , if no new answer is produced , @xmath24 is resolved by using answers after its state and all its dependent subgoals states are set to _
complete_. notice that a top - most looping subgoal does not return any answers until it is complete .
if @xmath24 is a looping subgoal but not a top - most one , @xmath24 will be resolved by using answers after its state is set to _
evaluated_. notice that @xmath24 s state can not be set to _ complete _ since @xmath24 is contained in a loop whose top - most subgoal has not been completely evaluated . for example , in figure [ fig : loops ] , q reaches its fixpoint only after the top - most looping subgoal p reaches its fixpoint .
as described in the definition of @xmath32 , an _ evaluated _ subgoal is never evaluated using rules again in the same round .
this optimization is called _ subgoal optimization _ in @xcite .
if evaluating a subgoal produces some new answers then the top - most looping subgoal will be re - evaluated and so will the subgoal ; and if evaluating a subgoal does not produce any new answer , then evaluating it again in the same round would not produce any new answers either .
therefore , the subgoal optimization is safe .
consider the following program , where p/2 is tabled , and the query p(a , y0 ) . ....
p(x , y):-p(x , z),e(z , y),memo(p(x , y ) ) .
( p1 ) p(x , y):-e(x , y),memo(p(x , y ) ) .
( p2 ) p(x , y):-check_completion(p(x , y ) ) .
( p3 ) e(a , b ) .
e(b , c ) .
.... the following shows the steps that lead to the production of the first answer : aa = aaa = aaa = aaa = aaa = aaa = aaa 1 : + @xmath36apply p1 + 2 : , e(z1,y0),memo(p(a , y0 ) ) + + 1 : + @xmath36 apply p2 + 3 : , memo(p(a , y0 ) ) + @xmath36 apply e(a , b ) + 4 : + @xmath36 add answer p(a , b ) after the answer p(a , b ) is added into the table , memo(p(a , b ) ) fails .
the failure forces execution to backtrack to p(a , y0 ) . aa = aaa = aaa = aaa = aaa = aaa = aaa 1 : + @xmath36 apply p3 + 5 : since p(a , y0 ) is a top - most looping subgoal which has not been completely evaluated yet , check_completion(p(a , y0 ) )
does not consume the answer in the table but instead starts re - evaluation of the subgoal .
aa = aaa = aaa = aaa = aaa = aaa = aaa 1 : + @xmath36apply p1 + 6 : , e(z1,y0),memo(p(a , y0 ) ) + @xmath36use answer p(a , b ) + 7 : , memo(p(a , y0 ) ) + @xmath36apply e(b , c ) + 8 : when the follower p(a , z1 ) is encountered this time , it consumes the answer p(a , b ) .
the current path leads to the second answer p(a , c ) . on backtracking ,
the goal numbered 6 becomes the current goal .
aa = aaa = aaa = aaa = aaa = aaa = aaa 6 : , e(z1,y0),memo(p(a , y0 ) ) + @xmath36use answer p(a , c ) + 9 : , memo(p(a , y0 ) ) goal 9 fails .
execution backtracks to the top goal and tries the clause p3 on it .
aa = aaa = aaa = aaa = aaa = aaa = aaa 1 : + @xmath36 apply p3 + 10 : since the new answer p(a , c ) is produced in the last round , the top - most looping subgoal p(a , y0 ) needs to be evaluated again .
the next round produces no new answer and thus the subgoal s state is set to _ complete_. after that the top - most subgoal returns the answers p(a , b ) and p(a , c ) . under the lazy strategy ,
answers are not returned immediately after they are produced but are returned via the table after all clauses are tried .
no answer is returned for a top - most looping subgoal until the subgoal is complete .
all loops are guaranteed to be real : for any loop @xmath37 where @xmath24 and @xmath25 are variants , @xmath24 must be an ancestor of @xmath25 . because each cluster of inter - dependent subgoals is completely evaluated before any answers are returned to outside of the cluster , the lazy strategy has good locality and is thus suited for finding all solutions .
for example , when the subgoal @xmath29 is encountered in the goal `` p(x),p(y ) '' , the subtree for p(x ) must have been explored completely and thus needs not be saved for evaluating p(y ) .
the cut operator can not be handled efficiently under the lazy strategy .
the goal `` @xmath38 '' produces all the answers for @xmath30 even though only one is needed .
the eager strategy prefers answer consumption and return over production . for a pioneer ,
answers are used first and rules are used only after all available answers are exhausted , and moreover a new answer is returned to its parent immediately after it is added into the table .
the following describes how the three primitives behave under the eager strategy . just as in the lazy strategy , @xmath24 is registered if it is not registered yet .
@xmath24 is resolved by using the tabled answers if @xmath24 is complete or @xmath24 is a follower of some former variant subgoal .
if @xmath24 is a pioneer , being encountered for the first time in the current round , it is resolved by using answers first , and then rules after all existing answers are exhausted .
if the answer @xmath24 is already in the table , then this primitive fails ; otherwise , this primitive succeeds after adding the answer @xmath24 into the table .
notice that @xmath24 is returned immediately after it is added into the table .
if @xmath24 is not new , then it must have been returned before .
if @xmath24 is a top - most looping subgoal , just as in the lazy strategy , we check whether any new answers are produced during the last iteration of @xmath24 .
if so , @xmath24 is evaluated again by calling @xmath32 .
otherwise , if no new answer is produced , this primitive fails after @xmath24 s and all its dependent subgoals states are set to _
complete_. if @xmath24 is a looping subgoal but not a top - most one , this primitive fails after @xmath24 s state is set to _
evaluated_. an _ evaluated _ subgoal is never evaluated using rules again in the same round .
notice that unlike under the lazy strategy , the primitive @xmath35 never returns any answers under the eager strategy . as described above , all the available answers must have been returned by @xmath32 and @xmath34 by the time @xmath35 is executed .
because of the need to re - evaluate a top - most looping subgoal , redundant solutions may be observed for a query .
consider , for example , the following program and the query `` p(x),p(y ) '' . ....
p(1):-memo(p(1 ) ) .
( r1 ) p(2):-memo(p(2 ) ) .
( r2 ) p(x):-check_completion(p(x ) ) .
( r3 ) .... the following derivation steps lead to the return of the first solution ( 1,1 ) for ( x , y ) .
aa = aaa = aaa = aaa = aaa = aaa = aaa 1 : , p(y ) + @xmath36 use r1 + 2 : , p(y ) + @xmath36 add answer p(1 ) + 3 : + @xmath36 loop found , use answer p(1 ) + when the subgoal p(y ) is encountered , it is treated as a follower and is resolved using the tabled answer p(1 ) .
after that the first solution ( 1,1 ) is returned to the top query . when execution backtracks to p(y ) , it fails since it is a follower and no more answer is available in the table .
execution backtracks to p(x ) , which produces and adds the second answer p(2 ) into the table .
aa = aaa = aaa = aaa = aaa = aaa = aaa 1 : , p(y ) + @xmath36 use r2 + 4 : , p(y ) + @xmath36 add answer p(2 ) + 5 : + @xmath36 use answer p(1 ) + when p(y ) is encountered this time , there are two answers p(1 ) and p(2 ) in the table .
so the next two solutions returned are ( 2,1 ) and ( 2,2 ) . when execution backtracks to goal 1 , the dummy ending rule is applied .
aa = aaa = aaa = aaa = aaa = aaa = aaa 1 : , p(y ) + @xmath36 use r3 + 6 : , p(y ) + since new answers are added into the table during this round , the subgoal p(x ) needs to be evaluated again , first using answers and then using rules .
the second round produces no answer but returns the four solutions ( 1,1 ) , ( 1,2 ) , ( 2,1 ) and ( 2,2 ) among which only ( 1,2 ) has not been observed before .
since answers are returned eagerly , a pioneer and a follower may not have an ancestor - descendant relationship .
because of the existence of this kind of _ fake _ loops and the necessity of iterating the evaluation of top - most looping subgoals , redundant solutions may be observed . in the previous example , the solutions ( 1,1 ) , ( 2,1 ) and ( 2,2 ) are each observed twice .
provided that the top - most looping subgoal p(x ) did not return the answer p(1 ) again in the second round , the solution ( 1,2 ) would have been lost .
the eager strategy is more suited than the lazy strategy for single - solution search . for certain applications such as planning
it is unreasonable to find all answers either because the set is infinite or because only one answer is needed . for these applications
the eager strategy is more effective than the lazy one .
cuts are handled more efficiently under the eager strategy .
the basic linear tabling framework described in the previous section does not distinguish between new and old answers .
the problem with this naive method is that it redundantly joins answers of subgoals that have been joined in early rounds .
semi - naive optimization @xcite reduces the redundancy by ensuring that at least one new answer is involved in the join of the answers for each rule . in this section ,
we introduce semi - naive optimization into linear tabling and identify sufficient conditions for it to be complete .
we also propose a technique called _ early answer promotion _ to further avoid redundant consumption of answers .
this optimization works with both the lazy and eager strategies . to make semi - naive optimization possible ,
we divide the answer table for each tabled subgoal into three regions : the names of the regions indicate the rounds during which the answers in the regions are produced : _ old _ means that the answers were produced before the previous round , _ previous _ the answers produced during the previous round , and _ current _ the answers produced in the current round .
the answers stored in _
previous _ and _ current _ are said to be _
new_. before each round is started , answers are promoted accordingly : _ previous _ answers become _ old _ and _ current _ answers become _
previous_. in our optimization , answers are consumed _
sequentially_. for a subgoal , either all the available answers or only new answers are consumed .
this is unlike in bottom - up evaluation where answers are consumed _ incrementally _ , i.e. , answers produced in a round are not consumed until the next round .
as will be discussed later , incremental consumption of answers as is done in bottom - up evaluation does avoid certain redundant joins but does not fit linear tabling since it may require more rounds to reach fixpoints .
a predicate @xmath27 _ calls _ a predicate @xmath39 if : ( 1 ) if @xmath39 occurs in the body of at least one rule in the definition of @xmath27 ( @xmath27 calls @xmath39 _ directly _ ) ; or ( 2 ) @xmath39 does not occur in the body of any rule in the definition of @xmath27 but there exists a predicate in the body of a rule in the definition of @xmath27 that calls @xmath39 ( @xmath27 calls @xmath39 _ indirectly _ ) .
the calling relationship constitutes a graph called a _ call graph_. for a given program , we find a level mapping from the predicate symbols in the program to the set of integers to represent the _ call graph _ of the program .
let @xmath40 be a level mapping .
we extend the notation to assume that @xmath41 for any subgoal @xmath42 of arity @xmath43 . for a given program ,
a level mapping @xmath40 represents the _ call graph _ if : for each rule `` @xmath2@xmath44@xmath45@xmath46 '' in the program , @xmath47 iff the predicate of @xmath48 does not call ( either directly or indirectly ) the predicate of @xmath2 , and @xmath49 iff the predicates of @xmath2 and @xmath48 call each other .
the level mapping as defined divides predicates in a program into several strata .
the predicate at each stratum depends only on those on the lower strata .
the level mapping is an abstract representation of the dependence relationship of the subgoals that may occur in execution .
if two subgoals @xmath24 and @xmath25 occur in a loop , then it is guaranteed that @xmath50 .
let `` @xmath2@xmath44@xmath45@xmath51 '' be a rule in a program and @xmath40 be the level mapping that represents the call graph of the program .
@xmath52 is called the _ last depending subgoal _ of the rule if @xmath53 and @xmath47 for @xmath54 .
the last depending subgoal @xmath52 is the last subgoal in the body that may depend on the head to become complete .
thus , when the rule is re - executed on a subgoal , all the subgoals to the right of @xmath52 that have occurred before must already be complete .
let `` @xmath2@xmath44@xmath45@xmath46 '' be a rule in a program and @xmath40 be a level mapping that represents the call graph of the program .
if there is no depending subgoal in the body , i.e. , @xmath47 for @xmath55 , then the rule is called a _
base rule_. let `` @xmath2@xmath44@xmath45@xmath51 '' be a rule where @xmath52 is the last depending tabled subgoal , and @xmath56 be a subgoal that is being resolved by using the rule in an iteration of a top - most looping subgoal @xmath57 .
for a combination of answers of @xmath8 , @xmath58 , and @xmath59 , if @xmath56 has occurred in an early round and the combination does not contain any new answers , then it is safe to let @xmath52 consume new answers only . because @xmath52 is the last depending subgoal ,
the subgoals @xmath60 , @xmath58 , and @xmath61 must have been completely evaluated when @xmath56 is re - evaluated .
let @xmath62 and @xmath63 be the _ old _ and _ new _ answers of the subgoal @xmath52 , respectively . for a combination of answers of @xmath8 , @xmath58 , and
@xmath59 , if the combination does not contain new answers then the join of the combination and @xmath62 must have been done and all possible answers for @xmath56 that can result from the join must have been produced during the previous round because the subgoal @xmath56 has been encountered before . therefore only new answers in @xmath63 should be used .
base rules need not be considered in the re - evaluation of any subgoals .
semi - naive optimization would be unsafe if it were applied to new subgoals that have never been encountered before .
the following example illustrates this possibility : .... ?
- p(x , y ) .
: -table p/2 .
p(x , y ) : - p(x , z),q(z , y ) .
( c1 ) p(b , c ) : - p(x , y ) .
( c2 ) p(a , b ) .
( c3 ) : -table q/2 .
q(c , d ) : - p(x , y),t(x , y ) .
( c4 ) t(a , b ) .
( c5 ) .... in the first round of p(x , y ) the answer p(a , b ) is added to the table by c3 , and in the second round the rule c2 produces the answer p(b , c ) by using the answer produced in the first round . in the third round ,
the rule c1 generates a new subgoal q(c , y ) after p(x , z ) consumes p(b , c ) .
if semi - naive optimization were applied to q(c , y ) , then the subgoal p(x , y ) in c4 could consume only the new answer p(b , c ) and the third answer p(b , d ) would be lost .
semi - naive optimization can lower the complexity of evaluation for some programs . consider the following example created by david s. warren : .... : -table p/2 .
p(x , y ) : - p(x , z),c(z , a , y ) .
p(x , y ) : - p(x , z),c(z , b , y ) .
p(x , x ) . .... which detects if a given string represented as facts c@xmath64
( @xmath65,@xmath66a or @xmath66b ) is a sentence of the regular expression @xmath67 . for a string @xmath68 , the query p(0,@xmath43 )
needs @xmath69 rounds to reach the fixpoint . with semi - naive optimization ,
the variants of p(x , z ) in the bodies consume only new answers , and therefore the program takes linear time . without semi - naive optimization , however , the program would take @xmath70 time since the variants of p(x , z ) would consume all existing answers . in our semi - naive optimization ,
answers produced in the current round are consumed immediately rather than postponed to the next round as in the bottom - up version , and answers are promoted each time a new round is started .
this way of consuming and promoting answers may cause certain redundancy . consider the conjunction @xmath71
. assume @xmath72 , @xmath73 , and @xmath74 are the sets of answers in the three regions ( respectively , _ old _ , _ previous _ , and _ current _ ) of the subgoal @xmath75 when @xmath75 is encountered in round @xmath76 .
assume also that @xmath0 had been complete before round @xmath76 and @xmath77 is the set of answers .
the join @xmath78 is computed for the conjunction in round @xmath76 .
assume @xmath79 , @xmath80 , and @xmath81 are the sets of answers in the three regions when @xmath75 is encountered in round i+1 . since answers are promoted before round @xmath82 is started , we have : aa = aaa = aaa = aaa = aaa = aaa = aaa @xmath83 + @xmath84 where @xmath85 denotes the new answers produced for @xmath75 after the conjunction @xmath71 in round @xmath76 . when the conjunction @xmath71 is encountered in round @xmath82 , the following join is computed .
aa = aaa = aaa = aaa = aaa = aaa = aaa @xmath86 notice that the join @xmath87 is computed in both round @xmath76 and @xmath82 .
we could allow last depending subgoals to consume answers incrementally as is done in bottom - up evaluation , but doing so may require more rounds to reach fixpoints .
consider the following example , which is the same as the one shown above but has a different ordering of clauses : .... ?
- p(x , y ) .
: -table p/2 .
p(a , b ) .
( c1 ) p(b , c ) : - p(x , y ) . ( c2 )
p(x , y ) : - p(x , z),q(z , y ) .
( c3 ) : -table q/2 .
q(c , d ) : - p(x , y),t(x , y ) . ( c4 )
t(a , b ) .
( c5 ) .... in the first round , c1 produces the answer p(a , b ) .
when c2 is executed , the subgoal in the body can not consume p(a , b ) since it is produced in the current round .
similarly , c3 produces no answer either . in the second round , p(a , b )
is moved to the _ previous _ region , and thus can be consumed .
c2 produces a new answer p(b , c ) .
when c3 is executed , no answer is produced since p(b , c ) can not be consumed . in the third round , p(a , b )
is moved to the _ old _ region , and p(b , c ) is moved to the _ previous _ region .
c3 produces the third answer p(b , d ) .
the fourth round produces no new answer and confirms the completion of the computation .
so in total four rounds are needed to compute the fixpoint .
if answers produced in the current round are consumed in the same round , then only two rounds are needed to reach the fixpoint . as discussed above
, sequential consumption of answers may cause redundant joins . in this subsection ,
we propose a technique called _ early promotion _ of answers to reduce the redundancy .
let @xmath75 be the first follower that exhausts its answers in the current round
. then all the answers of @xmath75 in the _ current _ region are promoted to the _ previous _ region once being consumed by @xmath75 .
consider again the conjunction @xmath71 where @xmath75 is the first follower that exhausts its answers .
the answers in the current region @xmath74 are promoted to the _ previous _ region after @xmath75 has consumed all its answers in round @xmath76 . by doing so
, the join @xmath87 will not be recomputed in round @xmath82 since @xmath74 must have been promoted to the _ old _ region in round @xmath82 .
consider , for example , the following program : .... ?
- p(x , y ) .
: -table p/2 .
p(a , b ) .
( c1 ) p(b , c ) : - p(x , y ) .
( c2 ) .... before c2 is executed in the first round , p(a , b ) is in the _ current _ region . executing c2
produces the second answer p(b , c ) . since the subgoal p(x , y ) in c2 is the first follower that exhausts its answers in the current round , it is qualified to promote its answers .
so the answers p(a , b ) and p(b , c ) are moved from the _ current _ region to the _ previous _ region immediately after being consumed by p(x , y ) . as a result
, the potential redundant consumption of these answers by p(x , y ) is avoided in the second round since they will all be transferred to the _ old _ region before the second round starts .
early promotion does not lose any answers .
first note that although answers are tabled in three disjoint regions , all tabled answers will be consumed except for some last depending subgoals that would skip the answers in their _ old _ regions ( see theorem 1 ) .
assume , on the contrary , that applying early promotion loses answers .
then there must be a last depending subgoal @xmath52 in a rule `` @xmath2@xmath44@xmath45@xmath51 '' and a tabled answer @xmath24 for @xmath52 such that @xmath24 has been moved to the _ old _ region before being consumed by @xmath52 so that @xmath24 will never be consumed by @xmath52 .
assume @xmath24 is produced in round @xmath76 by a variant of @xmath52 .
we distinguish between the following two cases : 1 .
the last depending subgoal @xmath52 is not selected in round @xmath76 . in round @xmath88 , @xmath52 is selected either because @xmath2 is new or some @xmath89 consumes a new answer . by theorem 1
, @xmath52 will consume all answers in the three regions , including the answer @xmath24 .
otherwise , @xmath24 must be produced by @xmath52 itself or a variant subgoal of @xmath52 that is selected either _ before _ or _ after _
@xmath52 in round @xmath76 .
if @xmath24 is produced by @xmath52 itself or _ before _ @xmath52 is selected , then the answer will be consumed by @xmath52 since promoted answers will remain new by the end of the round . if @xmath24 is produced by a variant _ after _
@xmath52 is selected , then the answer can not be promoted because @xmath52 exhausts its answers _ before _ the variant . in this case , the answer @xmath24 will remain new in the next round and will thus be consumed by @xmath52 . both of the above two cases contradict our assumption .
the proof then concludes .
changes to the prolog machine atoam @xcite are needed to implement linear tabling . in this section
we describe the changes to the data structures and the instruction set . to make the paper self - contained , we first give an overview of the atoam architecture .
the atoam uses all the data areas used by the wam .
the _ heap _ stores terms created during execution .
the register h points to the top of the heap .
the _ trail _ stack stores updates that must be undone upon backtracking .
the register t points to the top of the trail stack .
the _ control _
stack stores frames associated with predicate calls .
unlike in the wam where arguments are passed through argument registers , arguments in the atoam are passed through stack frames and only one frame is used for each predicate call .
each time a predicate is invoked by a call , a frame is placed on top of the local stack unless the frame currently at the top can be reused .
frames for different types of predicates have different structures . for standard prolog ,
a frame is either _ determinate _ or _ nondeterminate_. a nondeterminate frame is also called a
_ choice point_. the register ar points to the current frame and the register b points to the latest _
choice point_. a determinate frame has the following structure : & pointer to the parent frame + & continuation program pointer + & bottom of the frame + & top of the frame + & local variables + where btm points to the bottom of the frame , i.e. , the slot for the first argument , and top points to the top of the frame , i.e. , the slot just next to that for the last local variable .
the top register points to the next available slot on the stack .
the btm slot is not in the original version @xcite .
this slot is introduced for supporting garbage collection and co - routining .
the ar register points to the ar slot of the current frame .
arguments and local variables are accessed through offsets with respect to the ar slot .
an argument or a local variable is denoted as y(i ) where i is the offset .
arguments have positive offsets and local variables have negative offsets .
it is the caller s job to place the arguments and fill in the ar , and cp slots .
the callee fills in the btm and top slots and initializes the local variables .
a choice point frame contains , in addition to the slots in a determinate frame , four slots located between the top slot and local variables : & top of the heap + & top of the trail + & parent choice point + the cpf slot stores the program pointer to continue with when the current branch fails .
the slot h points to the top of the heap when the frame is allocated . as in the wam ,
a new register , called hb , is used as an alias for b->h .
when a variable is bound , it must be trailed if it is older than b or hb . a new data area , called _ table area _ , is introduced for memorizing tabled subgoals and their answers .
the _ subgoal table _ is a hash table that stores all the tabled subgoals encountered in execution . for each tabled subgoal and its variants
, there is an entry in the table , which is a record containing the following information : + + + + + the field subgoalcopy points to the copy of the subgoal in the table area . in the copy ,
all variables are numbered . therefore all variants of the subgoal are identical .
the field pioneerar points to the frame of the pioneer , which is needed for implementing cuts .
when the choice point of a tabled subgoal is cut off before the subgoal reaches completion , the field pioneerar will be set to null . when a variant of the subgoal is encountered again after
, the subgoal will be treated as a pioneer .
the field state indicates whether the subgoal is a looping subgoal , whether the answer table has been revised , and whether the subgoal is _ complete _ or _
evaluated_. when execution backtracks to a top - most looping subgoal , if the _ revised _ bit is set , then another round will be started for the subgoal .
a top - most looping subgoal becomes complete if this _ revised _ bit is unset after a round . at that time ,
the subgoal and all of its dependent subgoals will be set to _ complete_. as described in [ sub : evaluated ] , an _ evaluated _ subgoal is never evaluated again using rules in each round .
the topmostloopingsubgoal field points to the entry for the top - most looping subgoal , and the field dependentsubgoals stores the list of subgoals on which this subgoal depends . when a top - most looping subgoal becomes complete , all of its dependent subgoals turn to complete too .
the field answertable points to the answer table for this subgoal , which is also a hash table .
hash tables expand dynamically .
let g be the pointer to the record for a subgoal in the table .
the first answer in the answer table is referenced as ` g->answertable->firstanswer ` and the last answer is referenced as ` g->answertable->lastanswer ` . in the beginning , the answer table is empty and both firstanswer and lastanswer reference a dummy answer .
the frame for a tabled predicate contains the following two slots in addition to those slots stored in a choice point frame : + the subgoaltable points to the subgoal table entry , and the currentanswer points to the last answer that has been consumed .
the next answer can be reached from this reference on backtracking . when a frame is created , the slot currentanswer is initialized to be ` g->answertable->firstanswer ` where g is the pointer to the record for the tabled subgoal .
three new instructions , namely table_start , memo , and check_completion , are introduced into the atoam for encoding the three table primitives .
figure [ fig : ins ] shows the compiled code of an example program . ....
% : -tabled p/2 .
% p(x , y):-p(x , z),e(z , y ) .
% p(x , y):-e(x , y ) .
p/2 : table_start 2,1 fork r2 para_value
y(2 ) para_var y(-13 ) call p/2 % p(x , z ) para_value y(-13 ) para_value y(1 ) call e/2 % e(z , y ) memo r2 : fork r3 para_value y(2 ) para_value y(1 ) call e/2 % e(x , y ) memo r3 : check_completion p/2 .... the table_start instruction takes two operands : the arity ( 2 ) and the number of local variables ( 1 ) .
the fork instruction sets the cpf slot to hold the address to backtrack to on failure .
the parameter passing instructions ( para_value and para_var in this example ) pass arguments to the callee s frame .
the memo instruction is executed after an answer has been found .
the check_completion instruction takes the entrance ( p/2 ) as an operand so that the predicate can be re - entered when it needs re - evaluation . to implement semi - naive optimization
, we add the following two pointers into the record for each tabled subgoal : + where the pointer lastoldanswer points to the last answer in the old region and the pointer lastprevanswer points to the last answer in the previous region .
the check_completion instruction resets the pointers for all the tabled subgoals in the current cluster before it starts the next round : .... for each subgoal g in the current cluster { g->lastoldanswer = g->lastprevanswer ; g->lastprevanswer = g->answertable->lastanswer ; } .... the memo instruction is changed so that early promotion of answers is performed if the condition for promotion is met .
let g be the pointer to the tabled subgoal .
if the subgoal has exhausted all its answers in the table and early promotion has never be done before on the subgoal in the same round , then answers in the current region are promoted to the previous region : .... g->lastprevanswer = g->answertable->lastanswer .... the promoted answers will be moved to the old region before the start of the next round .
a bit vector is added into the frame for each tabled predicate to indicate if any new answer has been consumed by any tabled subgoal .
semi - naive optimization can be applied only if no tabled subgoal in the predicate has consumed any new answer . a new instruction , called last_depending_tabled_call ,
is introduced to encode last depending tabled subgoals . in the example shown in figure [ fig : ins ] , the `` call p/2 '' instruction is changed to `` last_depending_tabled_call p/2 '' to enable semi - naive optimization .
the last_depending_tabled_call instruction has the same behavior as the call instruction , but the callee can check the type of the instruction to see if it is invoked by a last depending tabled subgoal .
let g be the pointer to the current tabled subgoal .
the table_start instruction sets the currentanswer slot of the frame to ` g->lastoldanswer ` so that the subgoal consumes only new answers if : ( 1 ) the parent frame is a tabled frame ; ( 2 ) no bit in the bit vector in the parent frame is set , which means that no tabled subgoal has consumed any new answer ; and ( 3 ) the predicate is invoked by a last_depending_tabled_call instruction .
if any of these condition is not satisfied , the currentanswer slot is set to ` g->answertable->firstanswer ` and all the answers will be consumed by the subgoal .
we empirically compared the two answer consumption strategies and evaluated the effectiveness of semi - naive optimization .
we also compared the performance of b - prolog ( version 6.9 ) with xsb ( version 3.0 ) . a linux machine with 750mhz intel process and 512 gb ram was used in the experiment .
benchmarks from three different sources were used : datalog programs shown in figure [ fig : datalog ] with randomly generated graphs ; the chat benchmark suite @xcite ; and a parser , called _ atr _
, for the japanese language defined by a grammar of over 860 rules @xcite .
this section presents the experimental results and reports the statistics to support the results .
this section also gives experimental results on the warren s example for which slg as implemented in xsb has lower time complexity than linear tabling when semi - naive optimization ceases to be effective .
aa = aaa = aaa = aaa = aaa = aaa = aaa tcl : + + + tcr : + + + tcn : + + + sg : + table [ tab : strategies ] compares the two answer - consumption strategies in terms of speed and stack space efficiencies .
the difference of these two strategies in terms of cpu time is small on average .
this result implies that for programs with cuts declaring the use of the eager strategy would not cause significant slow - down .
the difference in the usage of stack space is more significant than in cpu time .
this is because , as discussed before , the lazy strategy has better locality than the eager strategy . &
lazy & eager & lazy & eager + tcl & 1 & 1.02 & 1 & 1.00 + tcr & 1 & 0.96 & 1 & 1.00 + tcn & 1 & 0.90 & 1 & 1.00 + sg & 1 & 0.89 & 1 & 1.02 + cs_o & 1 & 1.17 & 1 & 1.36 + cs_r & 1 & 1.09 & 1 & 1.36 + disj & 1 & 1.06 & 1 & 1.41 + gabriel & 1 & 1.08 & 1 & 1.18 + kalah & 1 & 1.17 & 1 & 2.03 + pg & 1 & 2.28 & 1 & 3.59 + peep & 1 & 0.99 & 1 & 2.88 + read & 1 & 0.85 & 1 & 2.22 + atr & 1 & 1.03 & 1 & 1.06 + @xmath90 & 1 & 1.12 & 1 & 1.62 + table [ tab : semi ] shows the effectiveness of semi - naive optimization in gaining speed - ups under both strategies . without this optimization ,
the system would consume over 30% more cpu time on average under either strategy .
our experiment also indicates that on average over 95% of the gains in speed are attributed to the _ early promotion _ technique .
& lazy & eager + tcl & 2.00 & 1.89 + tcr & 1.22 & 1.19 + tcn & 1.68 & 1.74 + sg & 1.22 & 1.51 + cs_o & 1.10 & 1.10 + cs_r & 1.09 & 1.10 + disj & 1.52 & 1.46 + gabriel & 1.32 & 1.15 + kalah & 1.52 & 1.41 + pg & 1.21 & 1.05 + peep & 1.09 & 1.11 + read & 1.98 & 1.27 + atr & 1.00 & 1.00 + @xmath90 & 1.38 & 1.31 + table [ tab : xsb ] compares bp with xsb on time and stack space efficiencies . for xsb ,
the stack space is the total of the maximum amounts of global , local , trail , choice point , and slg completion stack spaces .
the default setting , namely , the slg - wam and the local scheduling strategy , is used .
bp is faster than xsb on the datalog programs and the parser but slower than xsb on the chat benchmark suite ; and bp consumes considerably less stack space than xsb on some of the programs ( _ tcr _ , _ tcn _ , _ sg _ , and _ atr _ ) .
the results must be interpreted with two differences of the two compared systems taken into account : on the one hand , bp is on average more than twice as fast as xsb for standard prolog programs , and on the other hand the trie data structure used in xsb @xcite is far more advanced than hash tables used in bp for managing the table area .
it is unclear to what extent each difference contributes to the overall efficiency .
the yap implementation of slg - wam is up to twice as fast as xsb @xcite on the transitive closure and same - generation benchmarks with both chain and cyclic graphs .
this entails that the bp implementation of linear tabling is comparable in speed with the most sophisticated implementation of slg - wam for the datalog benchmarks . &
( lazy ) & cpu time & stack space + tcl & 1 & 1.85 & 0.81 + tcr & 1 & 1.46 & 33.41 + tcn & 1 & 1.31 & 32.84 + sg & 1 & 1.47 & 109.12 + cs_o & 1 & 0.37 & 0.57 + cs_r & 1 & 0.35 & 0.73 + disj & 1 & 0.68 & 0.82 + gabriel & 1 & 0.61 & 2.05 + kalah & 1 & 1.00 & 0.58 + pg & 1 & 0.76 & 1.85 + peep & 1 & 0.37 & 2.97 + read & 1 & 0.69 & 11.12 + atr & 1 & 2.26 & 21.24 + the empirical data on the usage of table space are not reported .
bp constantly consumes less table space than xsb for the benchmarks . in bp , both subgoal and answer tables
are maintained as dynamic hashtables . in xsb , in contrast , tables are maintained as tries @xcite .
the usage of table space is independent of the strategies and optimizations .
both bp and xsb would consume the same amount of table space if the same data structure were employed .
table [ tab : its ] reports the statistics on the maximum ( max its . ) and average ( ave . its . ) numbers of iterations for tabled subgoals to reach their fixpoints .
the column # subgoals shows the number of tabled subgoals . while for some programs , the maximum number of iterations performed is high ( e.g. , the maximum number for _ atr _ is 6 ) , the average numbers are quite low .
the necessity of re - evaluating looping subgoals has been blamed for the low speed of iteration - based tabling systems @xcite .
our new findings indicate that re - evaluation is not a dominant factor for the benchmarks .
this statistics well explain why an implementation of linear tabling could achieve comparable speed performance with slg - wam for the benchmarks .
tcl & 1 & 2 & 2.00 + tcr & 51 & 2 & 1.96 + tcn & 51 & 2 & 1.98 + sg & 153 & 2 & 1.32 + cs_o & 76 & 2 & 1.14 + cs_r & 76 & 2 & 1.16 + disj & 74 & 2 & 1.20 + gabriel & 59 & 2 & 1.20 + kalah & 102 & 3 & 1.24
+ pg & 48 & 2 & 1.13 + peep & 49 & 3 & 1.29 + read & 131 & 5 & 1.34 + atr & 7139 & 6 & 1.81 + the following is a slightly changed version of the warren s example which disenables semi - naive optimization : .... : -table p/2 .
p(x , y ) : - q(x , z),c(z , a , y ) .
p(x , y ) : - q(x , z),c(z , b , y ) .
p(x , x ) .
q(x , y ) : - p(x , y ) . .... since the last depending subgoals q(x , z ) in p/2 are not tabled , semi - naive optimization can not be applied to p/2 . for a string @xmath68 , the query p(0,@xmath43 )
needs @xmath69 iterations to reach the fixpoint .
since in each iteration the subgoal q(x , z ) is rewritten into p(x , z ) which returns all existing answers , the total time taken is @xmath70 .
in contrast , the program takes only @xmath91 time under slg . for the size n=5000
, it took bp 3.5 seconds to run the program while xsb only 15 milliseconds .
for the original version of the program to which semi - naive optimization is applicable , it took bp only 7 milliseconds .
there are three different tabling schemes , namely oldt and slg @xcite , cat @xcite , and iteration - based tabling including linear tabling @xcite and dra @xcite .
slg @xcite is a formalization based on oldt for computing well - founded semantics for general programs with negation .
the basic idea of using iterative deepening to compute fixpoints dates back to the et * algorithm @xcite . in slg - wam , a consumer fails after it exhausts all the existing answers and its state is preserved by freezing the stack so that it can be reactivated after new answers are generated .
the cat approach does not freeze the stack but instead copies the stack segments between the consumer and its producer into a separate area so that backtracking can be done normally .
the saved state is reinstalled after a new answer is generated .
chat @xcite is a hybrid approach that combines slg - wam and cat .
linear tabling relies on iterative computation of looping subgoals to compute fixpoints .
linear tabling is probably the easiest scheme to implement since no effort is needed to preserve states of consumers and the garbage collector can be kept untouched for tabling .
linear tabling is also the most space - efficient scheme since no extra space is needed to save states of consumers .
nevertheless , linear tabling without optimization could be computationally more expensive than the other two schemes .
the dra method @xcite is also iteration based , but it identifies looping clauses dynamically and iterates the execution of looping clauses to compute fixpoints . while in linear tabling iteration is performed on only top - most looping subgoals , in dra iteration is performed on every looping subgoal . in et * @xcite , every tabled subgoal is iterated even if it does not occur in a loop .
besides the difference in answer consumption strategies and optimizations , the linear tabling scheme described in this paper differs from the original version @xcite in that followers fail after they exhaust their answers rather than steal their pioneers choice points .
this strategy is originally adopted in the dra method .
the two consumption strategies have been compared in xsb @xcite as two scheduling strategies .
the lazy strategy is called _ local scheduling _ and the eager strategy is called _ single - stack scheduling_. another strategy , called _ batched scheduling _ , is similar to local scheduling but top - most looping subgoals do not have to wait until their clusters become complete to return answers .
their experimental results indicate that local scheduling constantly outperforms the other two strategies on stack space and can perform asymptotically better than the other two strategies on speed .
the superior performance of local scheduling is attributed to the saving of freezing stack segments .
although our experiment confirms the good space performance of the lazy strategy , it gives a counterintuitive result that the eager strategy is as fast as the lazy strategy .
this result implies that the cost of iterative evaluation is considerably smaller than that of freezing stack segments , and for predicates with cuts the eager strategy can be used without significant slow - down . in our tabling system
, different answer consumption strategies can be used for different predicates .
the tabling system described in @xcite also supports mixed strategies .
semi - naive optimization is a fundamental idea for reducing redundancy in bottom - up evaluation of logic database queries @xcite . as far as we know , its impact on top - down evaluation had been unknown before @xcite .
oldt @xcite and slg @xcite do not need this technique since it is not iterative and the underlying delaying mechanism successfully avoids the repetition of any derivation step .
an attempt has been made by guo and gupta @xcite to make incremental consumption of tabled answers possible in dra . in their scheme ,
answers are also divided into three regions but answers are consumed incrementally as in bottom - up evaluation . since no condition is given for the completeness and no experimental result is reported on the impact of the technique , we are unable to give a detailed comparison .
our semi - naive optimization differs from the bottom - up version in two major aspects : firstly , no differentiated rules are used . in the bottom - up version
differentiated rules are used to ensure that at least one new answer is involved in the join of answers for each rule .
consider , for example , the clause : aa = aaa = aaa = aaa = aaa = aaa = aaa @xmath92 the following two differentiated rules are used in the evaluation instead of the original one : aa = aaa = aaa = aaa = aaa = aaa = aaa @xmath93 + @xmath94 where @xmath95 denotes the new answers produced in the previous round for p. using differentiated rules in top - down evaluation can cause considerable redundancy , especially when the body of a clause contains non - tabled subgoals .
the second major difference between our semi - naive optimization and the bottom - up version is that answers in our method are consumed sequentially until exhaustion , not incrementally as in bottom - up evaluation .
a tabled subgoal consumes either all available answers or only new answers including answers produced in the current round .
neither incremental consumption nor sequential consumption seems satisfactory .
incremental consumption avoids redundant joins but may require more rounds to reach fixpoints .
in contrast , sequential consumption never need more rounds to reach fixpoints but may cause redundant joins of answers .
the early promotion technique alleviates the problem of sequential consumption . by promoting answers early from the _ current _ region to the _ previous _ region
, we can considerably reduce the redundancy in joins .
semi - naive optimization may lower time complexities in bottom - up evaluation @xcite .
the same result holds to the top - down version as demonstrated by warren s example .
our experimental results show that semi - naive optimization gives an average speed - up of over @xmath96 to linear tabling if answers are promoted early , and almost no speed gain if no answer is promoted early . in linear tabling ,
only looping subgoals need to be iteratively evaluated . for non - looping subgoals ,
no re - evaluation is necessary and thus semi - naive optimization has no effect at all on the performance .
most of the looping subgoals in our chosen benchmarks reach their fixpoints after 2 - 3 iterations . in general , more iterations are needed to reach fixpoints in bottom - up evaluation .
in addition , in bottom - up evaluation , the order of the joins can be optimized and no further joins are necessary once a participating set is known to be empty .
in contrast , in linear tabling joins are done in strictly chronological order . for a conjunction @xmath97 , the join @xmath98 is computed even if no answer is available for @xmath99 . because of all these factors , semi - naive optimization is not as effective in linear tabling as in bottom - up evaluation . our semi - naive optimization requires the identification of last depending subgoals . for this purpose ,
a level mapping is used to represent the call graph of a given program .
the use of a level mapping to identify optimizable subgoals is analogous to the idea used in the stratification - based methods for evaluating logic programs @xcite . in our level mapping ,
only predicate symbols are considered .
it is expected that more accurate approximations can be achieved if arguments are considered as well .
semi - naive optimization does not solve all the problems of recomputation in linear tabling .
recall the warren s example : .... : -table p/2 .
p(x , y ) : - p(x , z),c(z , a , y ) .
p(x , y ) : - p(x , z),c(z , b , y ) .
p(x , x ) .
.... assume there is a very costly non - tabled subgoal preceding p(x , z ) , then the subgoal has to be executed in each iteration even with semi - naive optimization .
this example demonstrates the acuteness of the problem of recomputation because the number of iterations needed to reach the fixpoint is not constant .
one treatment would be to table the subgoal to avoid recomputation , as suggested in @xcite , but tabling extra predicates can cause other problems such as over consumption of table space .
in this paper we have described two answer consumption strategies ( namely , _
lazy _ and _ eager _ strategies ) and semi - naive optimization for linear tabling .
we have compared the two strategies both qualitatively and quantitatively .
our results indicate that , while the lazy strategy has better space efficiency than the eager strategy , the eager strategy is comparable in speed with the lazy strategy .
this result implies that for all - solution search programs the lazy strategy should be adopted and for partial - solution search programs including programs with cuts the eager strategy should be used .
we have tailored semi - naive optimization to linear tabling and have given sufficient conditions for it to be complete .
moreover , we have proposed a technique called _ early answer promotion _ to reduce redundant consumption of answers .
our experimental result indicates that semi - naive optimization gives significant speed - ups to some programs .
linear tabling has several attractive advantages including its simplicity , ease of implementation , and good space efficiency .
early implementations of linear tabling were several times slower than xsb .
this paper has demonstrated for the first time that linear tabling with optimization is as competitive as slg on time efficiency as well for the benchmarks .
semi - naive optimization does not solve all the problems of recomputation in linear tabling .
there are programs for which recomputation can be costly , even leading to higher complexities .
the future work is to identify the patterns of such programs and find methods to deal with them .
the preliminary results of this article appear in acm ppdp03 and ppdp04 .
taisuke sato is supported in part by crest , and yi - dong shen is supported in part by the national natural science foundation of china grants numbered 60673103 and 60421001 .
\2001 . a simple scheme for implementing tabled logic programming systems based on dynamic reordering of alternatives . in _ proceedings international conference on logic programming ( iclp)_. lncs 2237 , 181195 . , nielson , h. r. , sun , h. , buchholtz , m. , hansen , r. r. , pilegaard , h. , and seidl , h. 2004 .
the succinct solver suite . in _ proc .
tools and algorithms for the construction and analysis of systems : 10th international conference ( tacas ) , lncs 2988_. 251265 . \1989 .
every logic program has a natural stratification and an iterated least fixed point model . in _
proceedings of the eighth acm sigact - sigmod - sigart symposium on principles of database systems_. acm press , 1121 . | recently there has been a growing interest of research in tabling in the logic programming community because of its usefulness in a variety of application domains including program analysis , parsing , deductive databases , theorem proving , model checking , and logic - based probabilistic learning .
the main idea of tabling is to memorize the answers to some subgoals and use the answers to resolve subsequent variant subgoals .
early resolution mechanisms proposed for tabling such as oldt and slg rely on suspension and resumption of subgoals to compute fixpoints .
recently , the iterative approach named linear tabling has received considerable attention because of its simplicity , ease of implementation , and good space efficiency .
linear tabling is a framework from which different methods can be derived based on the strategies used in handling looping subgoals .
one decision concerns when answers are consumed and returned . this paper describes two strategies , namely , _
lazy _ and _ eager _ strategies , and compares them both qualitatively and quantitatively .
the results indicate that , while the lazy strategy has good locality and is well suited for finding all solutions , the eager strategy is comparable in speed with the lazy strategy and is well suited for programs with cuts .
linear tabling relies on depth - first iterative deepening rather than suspension to compute fixpoints .
each cluster of inter - dependent subgoals as represented by a top - most looping subgoal is iteratively evaluated until no subgoal in it can produce any new answers .
naive re - evaluation of all looping subgoals , albeit simple , may be computationally unacceptable . in this paper , we also introduce semi - naive optimization , an effective technique employed in bottom - up evaluation of logic programs to avoid redundant joins of answers , into linear tabling . we give the conditions for the technique to be safe ( i.e. sound and complete ) and propose an optimization technique called
_ early answer promotion _ to enhance its effectiveness .
benchmarking in b - prolog demonstrates that with this optimization linear tabling compares favorably well in speed with the state - of - the - art implementation of slg .
# 1 [ firstpage ] prolog , semi - naive evaluation , recursion , tabling , memoization , linear tabling . |
consistency in preferences ( by pairwise comparisons ) can be traced to @xcite , published by kendall and smith in 1939 . by 1976 , there were four articles listed in @xcite , with the inconsistency in the title . in @xcite , the consistency index ( ci ) was introduced .
many other definitions followed .
all of them attempt to answer the most important question : how far have we departed from the consistent state which is well established by the consistency condition illustrated in fig . [ fig : fig2 ] in this text .
briefly , every cycle of three interrelated comparisons : @xmath1 , @xmath2 , and @xmath3 must be reducible to two comparisons , since the third comparison is a result of multiplication or division of these two comparisons .
for example , given @xmath1 and @xmath2 , @xmath3 should be equal to @xmath4 . if @xmath1 and @xmath3 are given , @xmath2 is equal to @xmath5 .
if @xmath2 and @xmath3 are given , @xmath1 is equal to @xmath6 . no other combination exists for a cycle of three comparisons .
if all three ratios are given , inconsistency in such cycle may take place and usually , it does .
earlier definitions of inconsistency were based on the total count of inconsistent triads .
evidently , such cardinal inconsistency was imprecise .
this study proposes to regard the inconsistency in pairwise as a degree or extent of disagreement modeled on the concept of probability measure .
our study also demonstrates a problem shown in @xcite , where the normalization was not used .
the interval @xmath0 $ ] ( or discrete values between 0 and 1 ) is used in many theories and situations including : * probability , * multi - valued logic ( as proposed by lukasiewicz ) , * fuzzy logic , * rough set theory , * theory of evidence . [ use01 ] there are also a number of zero - one laws of which the convergence of conditional expectations , known as `` lvy s zero-one law , '' can be summarized as `` if we are learning gradually all the information that determines the outcome of an event , then we will become gradually certain what the outcome will be , '' which has application to the consistency - driven pairwise comparisons ( recently used in @xcite , but based on inconsistency indicator introduced in @xcite ) .
the important interpretation of `` lvy s zero-one law '' is `` the gradual data gathering to determine the outcome of an event , then we will become gradually certain what the outcome will be '' since it has the direct application to the reduction of inconsistency in pairwise comparisons according to @xcite , p. lvy proved in 1937 ( see @xcite ) , that kolmogorov s theorem follows from a more general property of conditional probabilities , which says that + @xmath7 almost certainly equals 0 or 1 ( depending on whether @xmath8 is zero or not ) .
the use of interval @xmath0 $ ] for inconsistency indicators is not only coming from the occam s razor principle .
the section [ sec:3 ] shows that the lack of normalization in @xcite leads to problems which this study attempts to correct . in this study
, we will follow a business principle `` change a problem into opportunity '' and use it not only to correct @xcite but to reason for an important axiom .
the fallacy of the definition 3.6 on page 83 in @xcite : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ let @xmath9 be a group . a @xmath10-distance - based inconsistency indicator map ( in abbreviation : a @xmath10- inconsistency indicator map ) on the group @xmath11 is a function @xmath12 such that , for all @xmath13 , the following conditions are satisfied : 1
. @xmath14 ; 2 .
@xmath15 ; 3 .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ leads to a contradiction in the theory presented there .
the fallacy in the above definition is evidenced below by using the sequence of triads @xmath17 converging to a consistent matrix when the limit of @xmath18 , a natural number , is the infinity .
formally , an inconsistency indicator ( @xmath19 ) is defined for a pc matrix , say m. so , @xmath20 should be used .
however , a pc matrix : @xmath21\ ] ] is reduced to the three values : @xmath22 which generate the entire matrix . for this reason
, we will be using @xmath23 as a notation shortcut ( to save on typing and space ) .
to make our point , all we need to do is demonstrate just one counter - example ( for any particular distance ) .
in particular , our goal is achieved when we construct all triads in the sequence @xmath17 this way so that they have constant inconsistency greater than 0 ( e.g. , one or @xmath24 ) and yet such sequence converges .
in fact , it is enough to show the convergence for any sequence ( e.g. , cauchy sequence ) . evidently , @xmath17 is a cauchy sequence .
the well - established consistency condition for pc matrix @xmath25 $ ] is : @xmath26 these three pc matrix elements create a triad @xmath27 denoted here as @xmath28 .
a triad is consistent when its middle value is equal to the multiplication of the first and third value . to show that definition 3.6 in @xcite leads to contradiction ,
consider the following sequence of triads : @xmath29 the euclidean distance between the middle value in the triad @xmath17 and the result of two other values after the group operation ( multiplication ) is constant and equals to : @xmath30 .
the relative error is defined as : @xmath31 when the relative error definition , given by the equation ( [ eq:3 ] ) , is applied to the middle value of the triad @xmath17 , we get : @xmath32 ] to illustrate that the distance ( generalized or not ) can not serve as an inconsistency indicator , all we need to do is to produce just one counter - example .
[ fig : fig1 ] shows that `` one extra square '' added to a consistent triad @xmath33 ( hence `` one of two '' ) is worse than one extra square added to nine squares .
it is worthwhile to notice that the following triad : @xmath34 is simply reduced to : @xmath35 so , a triad may look like @xmath36 but @xmath36 is only consistent for @xmath37 , while @xmath38 if consistent for any @xmath39 ( technically , @xmath40 since pc matrix can not contain zero or negative values ) .
from the point of view of the proof of convergence of @xmath17 , it does not matter how the real limit looks like : @xmath38 or @xmath36 as long as the consistency condition holds for it .
replacing 1 in @xmath17 with any constant ( e.g. , @xmath24 which is a cosmic number ) just increases @xmath18 but the entire reasoning is unchanged . for the disbeliever in @xmath41 , we can use three equal sticks of the length 1 which ratios were inaccurately estimated to @xmath42 giving the relative error of @xmath43 , hence , 100% .
three other sticks , each 10 times shorter than the other stick , with the ratio estimations as @xmath44 give the 1% relative error while the euclidean distance is in both cases is identical and equal to 1 .
the sequence of triads @xmath17 converges to a consistent triad although each of its element has a constant inconsistency indicator making it inconsistent . defining the inconsistency indicator as an unnormalized distance leads to the following contradiction : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ a sequence of triads converges to a consistent triad
, however , each triad has ( according to the theory in @xcite ) the inconsistency indicator is equal to 1 for each triad in the sequence ; therefore , 1 has to be the sequence limit .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the fallacy in the definition of inconsistency in pairwise comparisons as distance for the multiplicative case in @xcite extends to the additive case as fig .
[ fig : diagram1 ] illustrates it .
areas of triangles are proportional to ratios they represent : 2 , 5 , and 3 .
the layout of these values resemble their location in a pc matrix given as : @xmath45 the first value in a triad @xmath28 represents a ratio of entities @xmath46 and @xmath47 : @xmath48 .
the middle triad value @xmath49 is located in the same matrix row as @xmath39 to the right of @xmath39 . the third value in the triad @xmath50 is located in the same column as @xmath51 below it .
values above the arrows represent the value for the third node to make the entire triad consistent .
for example , if the value of 2.5 above the upper arrow is used for the replacement of value 3 in the third node ( which this arrow does not connect ) , the triad becomes consistent ( since 2.5 * 2=5 ) .
similarly , using the value of 6 above the arrow connecting @xmath39 and @xmath51 for @xmath52 makes the triad @xmath28 consistent .
the above reasoning demonstrates why using the distance , which is not normalized , as inconsistency is unacceptable .
the distance between @xmath51 and @xmath53 in a triad @xmath28 remains constant ( e.g. 1 or @xmath24 ) for all triads in the sequence @xmath17 while the limit of the relative error of approximation is converging to 0 , hence , the triad becomes consistent at its limit with the expected inconsistency 0 .
it is worthwhile to notice that the relative ( hence , normalized ) error was needed to make the point but it can not be taken as a proof of the sufficiency .
the analogy to probability provides enough reasoning for the sufficiency .
the above reasoning applies to @xcite . in particular
, @xcite includes : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` definition 6.1 .
the consistency index of the matrix in equation 6.1 is given by @xmath54 ( 6.2 ) '' _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ which was analyzed above .
the similar reasoning applies to the additive case of pairwise comparisons , where ratio ( `` how many times ? '' ) is replaced by difference ( `` by how much ? '' ) against the use of the unnormalized distance as an inconsistency indicator .
[ fig : diagram1 ] illustrates the interpretation the inconsistency indicator as the area of the `` triangle '' which is reduced to the line for the consistent case .
it is easy to see that the inconsistency for a triad @xmath28 , defined in @xcite and finally simplified to @xmath55 in @xcite , does not suffer the above problem since the distance @xmath56 is normalized and converges to 0 .
it is worth mentioning that this may be an example for the need of normalization for applications since the distance normalization seems to be indispensable for the inconsistency indicator definition .
the need for a normalized rating scale for pairwise comparisons has been shown in @xcite .
the lack of the rating scale normalization leads to a paradox .
similarly , the use of `` central tendency '' in the inconsistency indicator definition leads to deviations as demonstrated in @xcite ( section 5 and 6 ) .
the two counter - examples and mathematical reasoning were provided that eigenvalue - based inconsistency , used in some customizations of pairwise comparisons , has no mathematical validity since it tolerates arbitrarily large errors when the size of a pc matrix increases to the infinity .
our reasoning can not be interpreted as `` normalization will fix it all '' since not every normalized distance can be used as inconsistency indicator , but whatever inconsistency indicator is defined , it has to be normalized for the reasons listed in the section [ sec:4 ] .
evidently , the list of reasons is , in all likeliness , incomplete , but sufficient to normalize all inconsistency indicators .
there is no question that we need to make risky moves in the absence of hard evidence and `` stipulative '' ( more relevant term : `` speculative '' ) definitions may be proposed .
however , such definitions should be clearly labeled so other researchers could be alerted and look for a validation or fallacy of such definitions .
the theory , based on stipulative foundations , which are not properly labeled , becomes hard to invalidate as time passes and the progress of science suffers .
one may only imagine that the eignevalue - based inconsistency in @xcite was `` stipulative '' since it was invalided in @xcite ) in 2014 .
definition 3.6 in @xcite could also be regarded as a stipulative definition .
however , it was not labeled ( or otherwise indicated ) as `` stipulative''in the text . the assumption that a generalized distance can be used as a map for inconsistency indicators was arbitrary ,
not rooted in the deep knowledge of pairwise comparisons , and unverified by real life examples . not only is it corrected by this study , but has facilitated the discovery of the need for normalization which is a paradigm shift in the research of the inconsistency in pairwise comparisons .
it is worth to notice that all coefficients of @xmath17 have the limit @xmath57 hence we may wonder whether or not such limit is attainable . observe that infinity is not a number and was rejected by the intuitionistic mathematics that was advocated by l.e.j .
as it turns , the existence of infinity can not be proved and it needs to be postulated ( something that many of us often overlook to do ) .
one of the philosophical issues of applied mathematics is the approximation theory of real numbers .
most transcendental real numbers can not be described since the set of transcendental real numbers is not countable , while a picture of one number after another can only show a countable set of numbers . evidently , at various degrees of knowledge and applications , the idealized framework of mathematics may be questioned by the the common sense with regards to the lack of effectiveness and/or feasibility of obtained results .
the normalization has become so popular ( especially in the applied science ) that it s impossible to trace its origins .
however , the probability is the high suspect , since it was already used ( in its embryonic and partially mystic way ) by cardano , who was a gambler , in the 14th century .
the probability has involved into one of the most powerful scientific methods universally used in most branches of applied and many theoretical sciences .
the normalization in probability theory is related to two of three axioms proposed by kolmogorov in 1933 hence it is of a considerable importance .
however , normalization might have been in use for a few centuries . according to @xcite ,
the acceptance of kolmogorov s axiomatization was not instantaneous but took tens of years ( some researchers still try to relax some of his axioms ) .
the issue of normalization leads to the non - trivial mathematical theory of boundedness and bornology in @xcite and in @xcite is related to compactification . a mathematical object @xmath46 such as a set or function
is _ bounded _ , provided that @xmath46 has a value such that all members of the set or function are less than this bound .
for example , the set of points in the half closed , half open interval @xmath58 are bounded above by 1 .
again , for example , the function @xmath59 is bounded above by 1 .
a _ bornology _ on a nonempty set @xmath11 is a collection @xmath60 of bounded sets @xmath47 that cover @xmath11 such that @xmath61 , and @xmath60 is stable under inclusions and finite unions .
the power set of a set @xmath11 ( denoted by @xmath62 ) is the collection of the subsets of @xmath11 , including @xmath11 .
the power set @xmath63 for @xmath64 is an example of a bornology .
boundedness and bornology are closely related to compactification .
the bordification is related to compatification .
a subset @xmath46 of a topological space @xmath11 is _ compact _ , provided , for every open cover of @xmath46 , there is a finite subcover of @xmath46 @xcite .
for example , every closed and bounded subset in @xmath65 is compact .
however , @xmath66 as well as any interval of @xmath66 is not compact .
for instance , let @xmath67 $ ] .
then @xmath63 is a topological space that is not compact , since @xmath63 has no finite subcover .
again , the set of all subsets of a geodetic line the plane is not compact for the same reason .
let @xmath11 be a non - compact space and let @xmath68 be a point that does not lie in @xmath11 .
the next step is to take a non - compact space @xmath11 and modify it to make it compact .
the result is a _ compactification _ of @xmath11 ( see , _ e.g. _ , @xcite ) .
one of the most important properties of normalization to @xmath0 $ ] is the stability under multiplication defined as : @xmath69 , t\cdot u \in [ 0,1]\ ] ] the maximality of @xmath0 $ ] is evident in terms of the upper bound , since we can not extend 0 to a negative value . if we extend 1 by the smallest positive value imaginable , say @xmath70 , the multiplication gives : @xmath71 it is worthwhile to notice that every non generative subinterval of @xmath0 $ ] is also stable under multiplication
. however , it is not maximal .
in other words , @xmath0 $ ] is the only bounded ( by two numbers excluding @xmath41 ) interval of the maximal size containing the multiplication of two non negative real numbers belonging to it .
the zero one infinity rule " is attributed to willem louis van der poel ( a pioneering dutch computer scientist , who is known for designing the zebra computer ) . by the above elimination of @xmath41 , our choice for the range of inconsistency indicators
is reduced to 0 and 1 .
acting in this spirit of the probability axioms , we postulate the lower and upper bounds for inconsistency indicator as : @xmath72 without the loss of generality .
unlike probability , inconsistency indicators do not need to reach the upper bound value .
it reflects the reality that no one can be indefinitely inconsistent with some exceptions such as `` 0 - 1 '' inconsistency indicator with 0 for all consistent triads and 1 otherwise .
evidently , @xmath73 for @xmath28 can not ever have a value of 1 since neither @xmath53 nor @xmath51 can be 0 ( @xmath22 are ratios hence must be strictly positive ) .
traditionally , most measures are constructed by setting two thresholds , namely , lower and upper bound .
the lower bound is nearly always ( if possible ) assumed as 0 .
the upper bound is @xmath74 , where @xmath75 for probability , fuzzy sets , rough sets , and more ; @xmath76 for various indicators such as school grades ; @xmath77 for percent values ( hence approximation error and business ) .
there are a number of reasons why the interval @xmath78 is not as good as @xmath79 $ ] but the most important reason is that the first half of @xmath78 is still @xmath78 . in practice ,
the inconsistency in pairwise comparisons is unavoidable unless we decide to use the simplified version of pairwise comparisons as described in @xcite with the minimum of @xmath80 comparisons of so - called principal generators which allow us to compute the rest of pc matrix elements from the consistency condition ( @xmath81 ) .
when inconsistency exists ( and it is nearly always as previously noticed ) , its degree is of great importance for the level of tolerance similar to @xmath82value in statistics .
the @xmath82value interpretation is the probability that , using a given statistical model , the statistical summary is less than or equal to the chosen significance level ( @xmath83 ) than the actual observed results ( e.g. , the difference between mean values of two compared groups ) .
needless to say that the value @xmath83 in statistics is arbitrarily assumed for applications . in medical research , @xmath84 is frequently assumed as a reasonable significance level . in pairwise comparisons , a similar level `` significance '' for the inconsistency .
it should be called as level of tolerance . using sn interval
@xmath0 $ ] is more convenient than @xmath78 .
for some inconsistency indicators ( e.g. , introduced in @xcite , called @xmath85 here ) , the level of tolerance is set to 1/3 since @xmath86 , and it does not have easy interpretation in @xmath78 .
the related concepts in statistics are confidence levels and confidence intervals introduced by neyman in 1937 in @xcite . for the inconsistency in pairwise comparisons , the issue of a tolerance level
is of as fundamental importance as the statistical significance in statistics .
the lack of `` point of reference '' in @xmath78 was a problem in @xcite , where a random inconsistency measure was introduced .
it is in the direct contradiction of stevens observations in @xcite stating that : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ `` the assignment can be any consistent rule .
the only rule not allowed would be random assignment , for randomness amounts in effect to a nonrule '' .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the interval @xmath0 $ ] is used for most indicators and measures in applied sciences as well as in theoretical sciences .
keeping the values in the interval @xmath0 $ ] may be the most natural way human brains process / assess concepts of nothing @xmath87 , a part of @xmath88 and total / sum ( which is 1 ) .
however , the most important reasoning for us is the following celebrated mathematical theorem .
normalizing mappings is well researched and there are many functions mapping @xmath89 onto @xmath88 .
for example , generalized logistic function or curve , also known as richards curve , a gompertz curve or gompertz function , named after benjamin gompertz , a sigmoid function , and many other functions are normalized .
it is worth mentioning that the inconsistency indicator , proposed in @xcite in 1993 , is normalized .
recently , it was simplified in @xcite to : @xmath90 the equation ( [ eq:7 ] ) has an equivalent form @xmath91 + a different ( and a bit unusual ) type of normalization is needed to prevent a rating scale paradox ( addressed in @xcite ) .
+ the original title of this section was `` why @xmath0 $ ] ? ''
but , after intensive study , it was changed to reflect the reality in the applied science .
it is mostly pure mathematics and theoretical physics where @xmath41 reigns although the speed limit is observed even in the theoretical physics .
using @xmath41 in proofs is not only highly desirable but often necessary .
however , variables of the indicators type are easier to process when they are normalized .
for this reason , using [ 0,1 ] should be a rule .
such a rule should only be relaxed if there is a good reason not to follow it .
the best way is to avoid inconsistency as much as possible .
sometimes , it is possible as @xcite demonstrates .
a decision maker may choose to provide just @xmath92 pairwise comparisons , which must be `` principal generators '' ( e.g. , located just above or below the main diagonal and other strategic places specified in @xcite ) .
the remaining pc matrix elements can be reconstructed so the entire pc matrix is consistent . using more than @xmath92 pairwise comparisons may ( and usually does ) lead to inconsistency , which is hard to manage without the use of heuristics .
however , avoidance of inconsistency has a reflection in the use of artificial levees in the river management .
when the artificial levee is breached , the damage to properties and lives can be greater than flooding which occurs without an artificial levee constructed .
when we enter more data and properly analyze their inconsistency to reduce it , the expectations are that the end result is improved .
we hope that the @xmath0 $ ] normalization of inconsistency indicators in pairwise comparisons may become one of `` the best practices '' of future knowledge management standards . in conclusion
, we have learned an important lesson here : when we use applied mathematics , we no longer have a complete freedom of using any term in any way we wish .
the mathematical distance can not serve as the inconsistence indicator as we demonstrated here by the use of the relative error .
the research has been supported in part by : * the polish ministry of science and higher education * the ministry of education , youth and sports czech republic within the institutional support for long - term development of a research organization in 2017 .
00 koczkodaj , w.w . ,
pairwise comparisons rating scale paradox , transactions on computational collective intelligence xxii , 9655:19 , 2016 .
koczkodaj , w. kulakowski , ligeza , a. , on the quality evaluation of scientific entities in poland supported by consistency - driven pairwise comparisons method , scientometrics , 99:911926 , 2014 .
koczkodaj , w.w . ; szybowski , j. ; wajch , e. , inconsistency indicator maps on groups for pairwise comparisons , international journal of approximate reasoning , 69 : 81-90 , 2016 .
krantz , s.g .
, essentials of topology with applications .
crc press , london , uk , 2016 .
j. neyman , outline of a theory of statistical estimation based on the classical theory of probability , philosophical transactions of the royal society of london .
series a , mathematical and physical sciences vol . 236 , no . 767 ( aug . 30 , 1937 ) , | in this study , we provide mathematical and practice - driven justification for using @xmath0 $ ] normalization of inconsistency indicators in pairwise comparisons . the need for normalization , as well as problems with the lack of normalization ,
are presented .
a new type of paradox of infinity is described .
keywords : pairwise comparisons , , inconsistency indicator , , normalization , , decision making . |
to study biological ion channels , the poisson - nernst - planck system ( , , @xcite , ) is commonly used as a model to describe the ionic flows in open ion channels .
the poisson - nernst - planck system is given by @xmath0,\\ \vspace{3 mm } \frac{\displaystyle \partial c_p}{\displaystyle\partial t}=\frac{\displaystyle\partial}{\displaystyle\partial x}\,\left[d_p\,(\frac{\displaystyle\partial c_p}{\displaystyle\partial x}+z_p\,c_p\,\frac{\displaystyle\partial \phi}{\displaystyle\partial x})\,\right ] , \\
-\varepsilon\,\frac{\displaystyle\partial^2 \phi}{\displaystyle\partial x^2}=z_n\,c_n+z_p\,c_p-\rho , \end{cases}\ ] ] for @xmath1 and @xmath2 . here
@xmath3 is the density of anions and @xmath4 is the density of cations ; @xmath5 is the electrostatic potential ; @xmath6 and @xmath7 are the diffusion coefficients ; @xmath8 and @xmath9 are positive integers ; @xmath10 is the permanent charge density ; @xmath11 is a parameter .
next we introduce a new type of poisson - nernst - planck system by considering the steric effect ( or size effect ) which occurs due to the fact that each atom within a molecule occupies a certain amount of space : @xmath12 + \frac{\displaystyle 1}{\displaystyle a(x)}\,\frac{\displaystyle\partial}{\displaystyle\partial x}\,\left[\,a(x)\,(g_{nn}\,c_n\,\frac{\displaystyle\partial c_n}{\displaystyle\partial x}+g_{np}\,c_n\,\frac{\displaystyle\partial c_{p}}{\displaystyle\partial x})\,\right],\\ \vspace{3 mm } \frac{\displaystyle \partial c_p}{\displaystyle\partial t}=\frac{\displaystyle \delta}{\displaystyle a(x)}\,\frac{\displaystyle\partial}{\displaystyle\partial x}\,\left[\,a(x)\,(\frac{\displaystyle\partial c_p}{\displaystyle\partial x}+z_p\,c_p\,\frac{\displaystyle\partial \phi}{\displaystyle\partial x})\,\right]+ \frac{\displaystyle 1}{\displaystyle a(x)}\,\frac{\displaystyle\partial}{\displaystyle\partial x}\,\left[\,a(x)\,(g_{pp}\,c_p\,\frac{\displaystyle\partial c_p}{\displaystyle\partial x}+g_{np}\,c_p\,\frac{\displaystyle\partial c_{n}}{\displaystyle\partial x})\,\right ] , \\
-\frac{\displaystyle\varepsilon}{\displaystyle a(x)}\frac{\displaystyle\partial}{\displaystyle\partial x}\left ( a(x)\frac{\displaystyle\partial \phi}{\displaystyle\partial x } \right)=z_n\,c_n+z_p\,c_p , \end{cases}\ ] ] for @xmath1 and @xmath2 . here
@xmath3 is the density of anions and @xmath4 is the density of cations ; @xmath5 is the electrostatic potential ; @xmath13 is the cross - sectional area of the ion channel at position @xmath14 ; @xmath15 , @xmath16 and @xmath17 are positive constants ; @xmath8 and @xmath9 are positive integers ; @xmath18 and @xmath11 are two parameters . assuming constant cross - sectional area @xmath19 , without loss of generality ,
we consider the following two - component _ drift - diffusion system _ @xmath20 where @xmath21 and @xmath22 are densities of the two species @xmath23 and @xmath24 , which are assumed to be nonnegative functions ; @xmath5 is the electric potential ; @xmath25 and @xmath26 are diffusion rates . throughout this paper , @xmath27 , @xmath28 , @xmath29 , @xmath30 , @xmath31 and @xmath32
are positive constants ; @xmath33 and @xmath34 are negative constants , unless otherwise specified .
@xmath35 is a bounded domain with smooth boundary . when the domain under consideration is extended to the entire space @xmath36 ,
becomes @xmath37 we note that , in the absence of @xmath38 , @xmath39 , @xmath40 , and @xmath41 , reduces to the keller - segel type equations @xmath42 keller - segel model is a classical model in chemotaxis introduced by keller and segel ( @xcite ) .
for the last two decades , there has been considerable literature devoted to the keller - segel model . for the keller - segel model
, we refer to the textbook @xcite , review papers @xcite and references therein . in this paper , we are concerned with stationary solutions to and , i.e. with time - independent solutions to the following elliptic systems @xmath43 using the elementary fact that @xmath44 , the first and second equations in can be rewritten as @xmath45 it is readily seen that if we can find @xmath23 , @xmath24 and @xmath46 satisfying the _ algebraic equations _
@xmath47 where @xmath48 and @xmath49 are constants , then such @xmath23 , @xmath24 and @xmath46 automatically form a solution of .
a natural question arises as to whether _ any _ solution of also satisfies .
it will be shown in proposition [ prop : equivalence of algebraic equations and differential equations ] that the answer is indeed affirmative when certain appropriate boundary conditions are imposed on the solutions , i.e. @xmath50 and @xmath51 where @xmath52 and @xmath53 .
for instance , @xmath54 on @xmath55 , or the neumann boundary conditions @xmath56 as a consequence , our problem now turns to establishing the existence of solutions for the _ differential algebraic equations _ ( daes ) : @xmath57 to the best of the authors knowledge , this is the first time such dae approach has been employed to drift - diffusion systems such as . in section [ sec : two species equations ] , we investigate in theorem [ thm : existence of solutions to aes ] the existence and uniqueness of solutions @xmath58 and @xmath59 to under the following hypothesis : * @xmath60 . note that is a system of nonlinear algebraic equations for which _ explicit solutions _ expressed by the form @xmath58 and @xmath59 in general can not exist . due to @xmath61
however , the solution @xmath58 and @xmath59 of _ in implicit form _ can be given _
uniquely_. with the aid of theorem [ thm : existence of solutions to aes ] , we are finally led to the following semilinear poisson equation @xmath62 where @xmath63 . to establish existence of solutions of under the zero neumann boundary condition @xmath64
more delicate properties of the nonlinearity @xmath65 and the solution @xmath58 , @xmath59 are explored in proposition [ prop : alge eqns ] .
these important properties turn out to be essential in proving the existence of solutions of by the standard _
direct method _ in the calculus of variations .
[ thm : main h1 ] assume that @xmath61 holds and @xmath48 and @xmath49 be fixed .
coupled with the neumann boundary conditions has a solution @xmath66 .
the intuition behind the differential algebraic equation approach we use in obtaining theorem [ thm : main h1 ] is elementary .
however , the result is remarkable in that only @xmath61 is needed to ensure the existence of solutions to the elliptic system together with the neumann boundary conditions . on the other hand , under the hypothesis : * @xmath67 , admits _ triple solutions _ ( see section [ sec : triple solutions ] ) .
moreover , in this case there may exist infinitely many solutions for .
[ thm : main h2 ] assume that @xmath68 holds and @xmath48 and @xmath49 be fixed .
coupled with the neumann boundary conditions has either a @xmath69 solution or infinitely many discontinuous solutions @xmath70 .
to begin with , we show that under the boundary conditions and , every solution of also solves , as mentioned in introduction .
[ prop : equivalence of algebraic equations and differential equations ] the systems of pdes @xmath71 together with the boundary conditions @xmath72 and @xmath73 where @xmath52 and @xmath53 , is equivalent to the system of algebraic equations @xmath74 integration by parts leads to the desired result .
indeed , @xmath75 however , @xmath76=0 gives @xmath77 this shows that @xmath78 , and thus @xmath79 should be a constant independent of @xmath14 . in a similar manner , we can prove that @xmath80 is a constant independent of @xmath14 from @xmath81 .
the proof of the converse is trivial .
hence the proof is finished .
since we have established proposition [ prop : equivalence of algebraic equations and differential equations ] , it is now necessary to investigate existence and uniqueness of solutions to . in particular , the solution of interest takes the form @xmath82 .
when the zero neumann boundary conditions @xmath83 on @xmath55 are considered , it is easy to see that these boundary conditions lead to @xmath84 and @xmath85 on @xmath55 .
however , note that the converse is not true . since the first two equations of are in divergence form , the total charges are conserved : @xmath86 where @xmath87 is a solution of .
it follows from the divergence theorem and the poisson equation in that in the case of the homogeneous neumann boundary condition @xmath88 , we have @xmath89 which is equivalent to the _ electroneutrality condition_. here @xmath32 and @xmath34 stand for the valence numbers of ions @xmath23 and @xmath24 , respectively .
[ thm : existence of solutions to aes ] assume @xmath61
. then for any given @xmath90 , there exists a unique solution @xmath91 of @xmath92 moreover , has a unique solution @xmath93 which can be represented implicitly as @xmath82 and @xmath94 are @xmath95 functions .
prove that for any given @xmath96 , we first show that the system of algebraic equations has at least one solution @xmath91 . in other words , the two curves @xmath97 and @xmath98 on the @xmath24-@xmath23 plane has at least one intersection point . to see this
, we differentiate @xmath99 with respect to @xmath24 to obtain @xmath100 which is always strictly less than zero because of the term @xmath101 appearing in the first equation of and @xmath23 can not take nonpositive value in @xmath101
. therefore , the profile @xmath102 of @xmath97 on the @xmath24-@xmath23 plane has the following property : * as @xmath103 , @xmath104 ; * as @xmath105 , @xmath106 ; * @xmath102 is decreasing in @xmath107 . in a similar manner
, we have for the second equation @xmath108 of , @xmath109 the profile @xmath110 of @xmath98 on the @xmath24-@xmath23 plane enjoys the following property : * as @xmath111 , @xmath105 ; * as @xmath104 , @xmath112 ; * @xmath110 is decreasing in @xmath113 . as a consequence
, it follows from the property of the profiles of the two curves @xmath97 and @xmath98 , that the two curves in the first quadrant of the @xmath24-@xmath23 plane intersect at least once .
that is , given any @xmath90 , we can find at least one solution @xmath91 which satisfies .
we eliminate the possibility of non - uniqueness of solutions @xmath114 to for a given @xmath115 by contradiction .
suppose that , contrary to our claim , there exist in the first quadrant of the @xmath24-@xmath23 plane two distinct solutions @xmath116 and @xmath117 which satisfy for given @xmath90 .
denote by @xmath118 ( respectively , @xmath119 ) the slope of the curve @xmath97 ( respectively , @xmath98 ) at @xmath114 .
it is easy to observe that , @xmath120 without loss of generality , we may assume that @xmath121 and @xmath122 . according to the intermediate value theorem
, there exists @xmath123 for which @xmath124 , where @xmath125 lies between @xmath126 and @xmath127 while @xmath128 lies between @xmath129 and @xmath130 .
however , using and , @xmath131 leads to @xmath132 it turns out that the last equation is equivalent to @xmath133 which contradicts @xmath61 .
consequently , for given @xmath115 , uniqueness of solutions to follows . by applying the local implicit function theorem at each point
@xmath134 which satisfies , we obtain @xmath95 smoothness of the solution @xmath82 .
the proof is completed .
certain important properties of solutions to are investigated in the next proposition .
these properties include non - simultaneousness of vanishing @xmath23 and @xmath24 , monotonicity of @xmath58 and @xmath59 , and monotonicity and convexity of @xmath102 . an important consequence of these properties is , which turns out to play a crucial role in employing the direct method to establish existence of solutions for the semilinear poisson equation .
[ prop : alge eqns ] as in theorem [ thm : existence of solutions to aes ] , assume @xmath61 .
then is uniquely solvable by the implicit functions @xmath82 .
concerning the properties of @xmath82 , we have * * ( non - simultaneous vanishing of @xmath58 and @xmath59)*. for fixed @xmath135 and @xmath114 which solves , for @xmath136 one of the following holds : @xmath137 + @xmath138 + @xmath139 + it is equivalent to say : there exists @xmath136 so that @xmath140 and @xmath141 can not simultaneously be true ; * * ( monotonicity of @xmath58 and @xmath59 ) * @xmath142 and @xmath143 can be expressed in terms of @xmath144 and @xmath145 , i.e. @xmath146 + @xmath147 + also , @xmath148 and @xmath149 for @xmath115 , i.e. @xmath58 is monotonically decreasing in @xmath115 , while @xmath59 is monotonically increasing in @xmath115 .
furthermore , @xmath142 and @xmath143 are uniformly bounded for @xmath115 .
in addition , we have @xmath150 for some constant @xmath151 ; * * ( monotonicity and convexity of @xmath102 ) * @xmath23 can be expressed by @xmath102 .
moreover , @xmath152 which implies competition between @xmath23 and @xmath24 , and @xmath153 we prove @xmath154 by contradiction .
assume to the contrary that there exists @xmath136 such that @xmath140 and @xmath141 hold simultaneously .
as @xmath18 is sufficiently small , @xmath155 are even smaller and @xmath156 . on the other hand , @xmath157 and @xmath158 have opposite signs since @xmath159 and @xmath160 .
therefore , in either @xmath161 can not be balanced with @xmath157 or @xmath162 can not be balanced with @xmath158 .
either case leads to a contradiction . to prove @xmath163 , we first differentiate the two equations in one by one with respect to @xmath46 , and obtain two equations in which the unknowns can be viewed as @xmath142 and @xmath143 . solving them gives @xmath142 and @xmath143 as stated in @xmath163 .
due to @xmath61 , it is easy to see that @xmath148 and @xmath149 for @xmath115 . using @xmath154 , when @xmath164 , clearly it follows from and that @xmath165 and @xmath166 for @xmath115 . as for the other two cases @xmath167 and
@xmath168 , it suffices to consider one of them since they are symmetric .
suppose that @xmath167 . for @xmath169 , and
lead to @xmath170 and @xmath171 on the other hand , when @xmath172 , estimates of @xmath142 and @xmath143 similar to and are given by @xmath173 and @xmath174 as a consequence , we conclude that there exists a constant @xmath151 such thtat @xmath175 . multiplying the first equation in by @xmath176 and the second equation in by @xmath177 , we obtain two equations .
subtraction of the two resulting equations gives an equation in terms of @xmath23 and @xmath24 .
implicit differentiation then gives the desired result in @xmath178 .
theorem [ thm : existence of solutions to aes ] and proposition [ prop : alge eqns ] lead us to consider the neumann problem for the semilinear poisson equation , i.e. @xmath179 we are now in a position to prove theorem [ thm : main h1 ] .
first of all , it can be shown by means of proposition [ prop : alge eqns ] that @xmath180 satisfies @xmath181 and @xmath182 of theorem [ thm : existence of solns to poission eqns with disconti nonlinearity ] . upon using theorems [ thm : existence of solutions to aes ] and [ thm : existence of solns to poission eqns with disconti nonlinearity ]
, we establish the existence of .
applying proposition [ prop : equivalence of algebraic equations and differential equations ] completes the proof of theorem [ thm : main h1 ] .
for the case where the hypothesis @xmath68 is assumed we can also establish an existence theorem for . in particular , in this case admits a _ triple solution _ in the sense that for some @xmath184 , we can find _ three _ pairs of solutions @xmath185 @xmath186 .
see figure [ fig : u = u(phi ) ] for an example .
this is essentially different from the case where @xmath61 holds .
[ 1.0 ] as shown in figure [ fig : u = u(phi ) ] , there may exist a _
triple solution _ and thus uniqueness no longer holds when @xmath61 is violated , i.e. when @xmath67 .
when non - uniqueness occurs , it is readily seen that there must be a @xmath187 such that @xmath188 .
due to and , this can occur only when @xmath189 .
indeed , it is possible that the numerator @xmath190 ( @xmath191 ) in ( ) as @xmath192 ( @xmath193 ) . however , as we have mentioned in the proof of proposition [ prop : alge eqns ] that @xmath194 it is therefore necessary to find @xmath114 at which @xmath195 vanishes by solving the equations : @xmath196 it turns out that is reduced to a single equation @xmath197 by eliminating @xmath46 and substituting @xmath198 .
more precisely , @xmath199 now the question remains to determine the profile of @xmath197 . to this end , we first observe that @xmath200 makes sense only when @xmath201 , where @xmath202 .
also , it is readily verified that @xmath203 to find extreme points of @xmath197 , we calculate @xmath204 where @xmath205 , and @xmath206 we remark that the denominator of @xmath207 is always away from @xmath208 since @xmath201 . let the numerator of be zero .
then the four roots are @xmath209 , @xmath126 , @xmath127 , and @xmath210 , where @xmath126 , @xmath127 , and @xmath210 are the three roots of @xmath211 .
indeed , @xmath126 , @xmath127 , and @xmath210 are three distinct real roots . to show this , we use fan shengjin s method . as in @xcite , let @xmath212 and the _ discriminant _
[ lem : discriminant ] there are three possible cases using the discriminant @xmath214 : * if @xmath215 , then @xmath211 has one real root and two nonreal complex conjugate roots . * if @xmath216 , then @xmath211 has three real roots with one root which is at least of multiplicity 2 . * if @xmath217 , then @xmath211 has three distinct real roots .
it can be shown that @xmath218 can not be true under the assumption @xmath68 .
using lemma [ lem : discriminant ] , we conclude that the cubic equation @xmath219 has three distinct real roots @xmath220 .
due to @xmath221 and @xmath222 , it is easy to see that either @xmath223 or @xmath224 .
however , the case @xmath223 can not occur since @xmath200 makes sense only for @xmath225 .
for the other case @xmath224 , @xmath210 can not be a extreme point of @xmath200 because of @xmath226 .
accordingly , there are _ at most two _ extreme points @xmath126 and @xmath127 .
we have by the asymptotic behavior @xmath227 as @xmath228 or @xmath104 , which leads to the fact that the number of extreme points of @xmath197 can only be _
odd_. as a consequence , one of @xmath126 and @xmath127 can not be a extreme point of @xmath197 and the other one is _ the _ extreme point of @xmath197 .
suppose that @xmath127 is a extreme point of @xmath197 , then so is @xmath126 , which yields a contradiction .
therefore , @xmath126 is _ the _ extreme point of @xmath197 .
in fact , it can be shown that @xmath229 so that @xmath230 dose not make sense
. now a question remains , i.e. , how to determine the sign of @xmath231 ?
we see that the maximum of @xmath232 is attained at @xmath233 or @xmath234 .
moreover , the criterion for determining the roots of the equation @xmath197 is stated as : * when @xmath235 , @xmath197 has two distinct positive solutions ; * when @xmath236 , @xmath197 has no solutions ; * when @xmath237 , @xmath197 has a unique positive solution ( i.e. @xmath233 ) . noting that as @xmath237 and @xmath238 , we have @xmath239 evaluated at @xmath233 . in other words
@xmath144 and @xmath145 has a _ reflection point _ at @xmath240 for some @xmath241 . on the other hand , as @xmath236
, the denominator @xmath195 of ( also ) always keeps it sign since @xmath197 has no solutions .
indeed , it is easy to see that @xmath242 , which leads to @xmath148 and @xmath149 .
[ thm : g11g22<g12g21 existence of alge soln ] assume that @xmath68 holds and @xmath243 solves @xmath92 then * * ( triple solutions ) * when @xmath235 , there exist @xmath244 such that * * @xmath144 ( and @xmath145 ) takes three distinct values for @xmath245 ; * * @xmath144 ( and @xmath145 ) can be represented uniquely for @xmath246 ; * * @xmath144 ( and @xmath145 ) takes two distinct values at @xmath247 ; * * ( uniqueness of monotone solutions ) * when @xmath236 , @xmath144 and @xmath145 can be represented uniquely for @xmath115 .
moreover , @xmath148 and @xmath149 for @xmath115 ; * * ( unique and monotone solutions with inflection points ) * when @xmath237 , @xmath144 and @xmath145 can be represented uniquely for @xmath115
. furthermore , there exists @xmath241 such that * * @xmath148 and @xmath149 for @xmath248 ; * * @xmath249 at @xmath240 .
we are able to explain @xmath48 @xmath49 in the following manner . according to theorem [ thm : g11g22<g12g21 existence of alge soln ]
, the sign of @xmath231 determines the properties , such as the number of solutions , of . here
@xmath126 is the largest positive root of @xmath211 and the discriminant as shown in relies on @xmath250 and @xmath183 @xmath251 .
consequently , @xmath126 is determined once @xmath250 and @xmath183 @xmath251 are given .
now @xmath252 and @xmath253 @xmath251 can be suitably chosen such that anyone of the three cases as described in theorem [ thm : g11g22<g12g21 existence of alge soln ] can occur .
[ thm : existence of solns to poission eqns with disconti nonlinearity ] assume that @xmath35 is an open and bounded domain with smooth boundary @xmath55 and * @xmath254 is piecewise continuous and discontinuous at finite points for @xmath255 ; @xmath256 for all @xmath257 ; * there exist constants @xmath258 , such that @xmath259 for @xmath260 and @xmath261 for @xmath262
. then the following neumann problem @xmath263 has a unique weak solution @xmath264 . for @xmath265 ,
let @xmath266 as seen previously , it is sufficient to consider the minimizing problem @xmath267 using @xmath182 , we have @xmath268 therefore , there exists @xmath269 such that @xmath270 and we can find a sequence of functions @xmath271 with @xmath272 .
now we show that @xmath273 is coercive , that is @xmath274 where @xmath275 . indeed ,
when @xmath276 as @xmath277 , we clearly have . on the other hand , when @xmath278 is bounded and @xmath279 as @xmath277
, it follows that also holds . to see this , @xmath280 since
@xmath279 as @xmath277 and @xmath281 , we conclude that @xmath282 and @xmath283 as @xmath277 .
the coercivity condition leads to the fact that there exists a constant @xmath284 such that @xmath285 ( otherwise , the unboundness of @xmath286 together with contradicts ) .
it follows that the sequence @xmath287 is bounded in @xmath288 , and there exists a subsequence @xmath289 of @xmath290 with @xmath291 we select a subsequence @xmath292 of @xmath289 which converges in @xmath293 , i.e. @xmath294 since @xmath295 convergence implies pointwise convergence of a subsequence almost everywhere , we have @xmath296 where @xmath297 is a subsequence of @xmath298 .
we denote this subsequence obtained by @xmath299 . to find a minimizer of our minimizing problem
, it suffices to establish weak lower semicontinuity , i.e. @xmath300 which gives @xmath301 this shows that @xmath125 solves the minimizing problem .
now we prove .
indeed , @xmath302 is continuous since @xmath303 for all @xmath257 .
since @xmath304 a.e . in @xmath305 as @xmath306 , we have @xmath307 a.e .. applying fatou s lemma to obtain @xmath308 yields @xmath309 where we have used the fact that @xmath310 as @xmath306 . in fact , elliptic regularity of weak solutions ensure smoothness of @xmath125 .
this completes the proof of the theorem .
in this section , the following initial conditions @xmath311 are imposed to determine completely the evolution . since the first two equations of are in divergence form ,
total charges are conserved : @xmath312 the means @xmath313 and @xmath314 of the charges @xmath23 and @xmath24 over @xmath305 are defined by @xmath315 respectively .
the following two lemmas are crucial in proving theorem [ thm : toward the equilibria ] .
we note that csiszr - kullback - pinsker inequality ( references refer to the ones cited in ) comes from information theory .
[ lem : logarithmic sobolev inequality ] for @xmath320 , @xmath321 , @xmath322 where @xmath323 , @xmath324 , @xmath325 is the logarithmic sobolev constant , and @xmath326 is a constant depending on @xmath327 and @xmath328 . under certain hypotheses on the initial conditions and the parameters appearing in ,
we show that the solution of the initial - boundary value problem tends to @xmath313 and @xmath313 in the @xmath333 sense . more precisely , we have in @xmath338 and @xmath339 , @xmath340 is the poincar constant .
if a global - in - time solution @xmath87 of the time - varying problem , , exists , then it tends to the corresponding constant steady - state solution @xmath341 in the @xmath342-norm as @xmath343 , i.e. @xmath344 furthermore , the time - varying solution @xmath87 converges in the @xmath342-norm exponentially fast to its mean with explicit rate @xmath345 : @xmath346 where @xmath345 will be specified in the proof , and @xmath347 is the constant in the csiszr - kullback - pinsker inequality ( lemma [ lem : csiszar - kullback inequality ] ) . as in @xcite ,
we define the _ relative entropy functional _ @xmath348(t)$ ] by @xmath349(t)=\int_{\omega}\ , \big(u\,\log \frac{u}{\bar{w}_1 } + v\,\log \frac{v}{\bar{w}_2 } \big)\,dx.\ ] ] by virtue of the elementary fact @xmath350 for @xmath351 , it can be shown that @xmath348(t)\geq0 $ ] for all @xmath2 . indeed , @xmath352(t)+
\int_{\omega}\,\bar{w}_1\,dx-\int_{\omega}\,u\,dx+ \int_{\omega}\,\bar{w}_2\,dx-\int_{\omega}\,v\,dx \notag\\ = & h[u , v](t ) . \end{aligned}\ ] ] note in particular that @xmath353 , @xmath354 for all @xmath2 . for simplicity of notation ,
let @xmath355 and @xmath356 .
we calculate the time derivative of @xmath348 $ ] to obtain @xmath357 = & \int_{\omega}\,u_t\,\big(\log \frac{u}{\bar{w}_1}+1\big ) + v_t\,\big(\log \frac{v}{\bar{w}_2}+1\big)\,dx\notag\\ = & \int_{\omega}\ , \nabla\cdot\mathcal{i}_1 \,\big(\log \frac{u}{\bar{w}_1}+1\big ) + \nabla\cdot\mathcal{i}_1 \,\big(\log \frac{v}{\bar{w}_2}+1\big)\,dx\notag\\ = & \int_{\omega}\ , -\mathcal{i}_1\cdot\nabla\big(\log \frac{u}{\bar{w}_1}+1\big ) -\mathcal{i}_1\cdot\nabla\big(\log \frac{v}{\bar{w}_2}+1\big)\,dx\notag\\ = & -\int_{\omega}\ , \big(\mathcal{i}_1\cdot \frac{\nabla u}{u}+ \mathcal{i}_2\cdot \frac{\nabla v}{v}\big)\,dx\notag\\ = & -\int_{\omega}\ , \big(d_1\,u^{-1}\,|\nabla u|^2+\vartheta_1\,\nabla\phi\cdot\nabla u+ g_{11}\,|\nabla u|^2+g_{12}\,\nabla v\cdot\nabla u \notag\\ & \hspace{14.0 mm } d_2\,v^{-1}\,|\nabla v|^2+\vartheta_2\,\nabla\phi\cdot\nabla v+ g_{22}\,|\nabla v|^2+g_{21}\,\nabla u\cdot\nabla v\big)\,dx\notag\\ = & -d_1\int_{\omega}\,u^{-1}\,|\nabla u|^2\,dx
-d_2\int_{\omega}\,v^{-1}\,|\nabla v|^2\,dx \notag\\ & -\gamma_1\,\vartheta_1 \int_{\omega}\,u^2\,dx -\gamma_2\,\vartheta_2
\int_{\omega}\,v^2\,dx -(\gamma_1\,\vartheta_2+\gamma_2\,\vartheta_1 ) \int_{\omega}\,u\,v\,dx\notag\\ & -g_{11}\int_{\omega}\ , |\nabla u|^2\,dx -g_{22}\int_{\omega}\ , |\nabla v|^2\,dx -(g_{12}+g_{21})\int_{\omega}\ , \nabla u\cdot\nabla v\,dx,\ ] ] where we have used green s first identity and @xmath358 to get @xmath359 as well as @xmath360 . on the other hand , it is readily seen that @xmath361 .
in particular , we have @xmath362 and @xmath363 . as a consequence ,
@xmath357 = & -d_1\int_{\omega}\,u^{-1}\,|\nabla u|^2\,dx -d_2\int_{\omega}\,v^{-1}\,|\nabla v|^2\,dx \notag\\ & -\gamma_1\,\vartheta_1 \int_{\omega}\,(u-\bar{w}_1)^2\,dx -\gamma_2\,\vartheta_2 \int_{\omega}\,(v-\bar{w}_2)^2\,dx \notag\\ & -(\gamma_1\,\vartheta_2+\gamma_2\,\vartheta_1 ) \int_{\omega}\,(u-\bar{w}_1)(v-\bar{w}_2)\,dx\notag\\ & -g_{11}\int_{\omega}\ ,
|\nabla u|^2\,dx -g_{22}\int_{\omega}\ , |\nabla v|^2\,dx -(g_{12}+g_{21})\int_{\omega}\ , \nabla u\cdot\nabla v\,dx- \mathcal{a}(\bar{w}_1,\bar{w}_1),\ ] ] where @xmath364 . under the hypotheses * ( @xmath3651 ) * and * ( @xmath3652 ) * , applying the cauchy inequality @xmath366 , the poincar inequality @xmath367 , where @xmath340 is the poincar constant , and the logarithmic sobolev inequality in lemma [ lem : logarithmic sobolev inequality ] , we find @xmath357 \leq & -4\,d_1\,c_l^{-1}\int_{\omega}\,u\,\log \frac{u}{\bar{w}_1}\,dx -4\,d_2\,c_l^{-1}\int_{\omega}\,v\,\log \frac{v}{\bar{w}_2}\,dx \notag\\ & -\gamma_1\,\vartheta_1 \int_{\omega}\,(u-\bar{w}_1)^2\,dx -\gamma_2\,\vartheta_2 \int_{\omega}\,(v-\bar{w}_2)^2\,dx \notag\\ & -\frac{1}{2}(\gamma_1\,\vartheta_2+\gamma_2\,\vartheta_1 ) \int_{\omega}\,(u-\bar{w}_1)^2\,dx -\frac{1}{2}(\gamma_1\,\vartheta_2+\gamma_2\,\vartheta_1 ) \int_{\omega}\,(v-\bar{w}_2)^2\,dx\notag\\ & -g_{11}\int_{\omega}\ , |\nabla u|^2\,dx -g_{22}\int_{\omega}\ , |\nabla v|^2\,dx\notag\\ & + \frac{1}{2}(g_{12}+g_{21})\int_{\omega}\ , |\nabla u|^2\,dx + \frac{1}{2}(g_{12}+g_{21})\int_{\omega}\ , |\nabla v|^2\,dx -\mathcal{a}(\bar{w}_1,\bar{w}_1)\notag\\ \leq & -4\,\min(d_1,d_2)\,c_l^{-1}\ , h[u , v ] -\frac{1}{2 } ( \gamma _ 2\,\vartheta_1+\gamma _ 1\,(2\,\vartheta_1+\vartheta _ 2))\,\int_{\omega}\,(u-\bar{w}_1)^2\,dx \notag\\ & -\frac{1}{2 } ( \gamma _ 1\,\vartheta_2+\gamma _ 2\,(2\,\vartheta_2+\vartheta _
1))\,\int_{\omega}\,(v-\bar{w}_2)^2\,dx -\mathcal{a}(\bar{w}_1,\bar{w}_1)\notag\\ & + \frac{1}{2}(g_{12}+g_{21}-2\,g_{11})\int_{\omega}\ , |\nabla u|^2\,dx + \frac{1}{2}(g_{12}+g_{21}-2\,g_{22})\int_{\omega}\ , |\nabla v|^2\,dx\notag\\ \leq & -4\,\min(d_1,d_2)\,c_l^{-1}\ , h[u , v ] -\frac{1}{2 } ( \gamma _ 2\,\vartheta_1+\gamma _ 1\,(2\,\vartheta_1+\vartheta _ 2))\,\int_{\omega}\,(u-\bar{w}_1)^2\,dx \notag\\ & -\frac{1}{2 } ( \gamma _ 1\,\vartheta_2+\gamma _ 2\,(2\,\vartheta_2+\vartheta _ 1))\,\int_{\omega}\,(v-\bar{w}_2)^2\,dx -\mathcal{a}(\bar{w}_1,\bar{w}_1)+\notag\\ & \frac{c_p^{-1}}{2}(g_{12}+g_{21}-2\,g_{11})\int_{\omega}\ , ( u-\bar{w}_1)^2\,dx + \frac{c_p^{-1}}{2}(g_{12}+g_{21}-2\,g_{22})\int_{\omega}\ , ( v-\bar{w}_2)^2\,dx\notag\\ \leq & -4\,\min(d_1,d_2)\,c_l^{-1}\ , h[u , v]-\mathcal{a}(\bar{w}_1,\bar{w}_1 ) . \ ] ] in fact , @xmath368 can be rewritten as @xmath369 due to the _ compatibility condition ( or the electroneutrality condition ) _ of the neumann problem for @xmath46 , we show that @xmath370 . to this end , the boundary condition @xmath88 on @xmath55 yields @xmath371 or @xmath372 , which results in @xmath370 .
this leads to the following initial value problem involving a differential inequality @xmath373(t)\leq -\mathcal{b}(d_1,d_2,c_l)\ , h[u , v](t ) , \quad t>0,\\ \hspace{4mm}h[u , v](0)=h_0 , \end{cases}\ ] ] where @xmath374 ; @xmath375(0)$ ] determined by is a positive constant depending on @xmath376 and @xmath377 .
gronwall s inequality yields @xmath378<h0 e - t } h[u , v](t)\leq e^{-\mathcal{b}\,t}\,h_0,\ ] ] where @xmath379 .
thanks to * ( @xmath3653 ) * , @xmath336 . since from the csiszr - kullback - pinsker inequality ( lemma [ lem : csiszar - kullback inequality ] )
, the @xmath342-norm of @xmath380 and @xmath381 can be controlled by @xmath348(t)$ ] . the same decay rate as in is given by @xmath382 this completes the proof of the theorem . | a method based on a differential algebraic equation ( dae ) approach is employed to find stationary solutions of the poisson - nernst - planck equations with steric effects ( pnp - steric equations ) . under appropriates
boundary conditions , the equivalence of the pnp - steric equations and a corresponding system of daes is shown . solving this system of daes leads to the existence of stationary solutions of pnp - steric equations .
we show that for suitable range of the parameters , the steric effect can produces infinitely many discontinuous stationary solutions . moreover under a stronger intra - species steric effect , we prove that a smooth solution converges to a constant stationary solution . |
accurate determination of the parton distribution function ( pdf ) inside proton is an essential part of analyzing data in deep - inelastic scattering ( dis ) processes .
precise measurements from high energy hadron colliders such as tevatron and large hadron collider ( lhc ) require the inclusion of higher order effects in proton - proton scattering .
it seems that the photon - induced drell - yan ( dy ) process such as @xmath3has a significant contribution ( @xmath4 ) to the dilepton invariant mass distribution .
recent results from high mass dy production in atlas [ showed that contribution of photon distribution inside proton has the same importance as the other different pdfs set . to calculate the cross section of such dy process
, one needs to know the photon distribution function inside proton , @xmath5 .
furthermore , because the lhc is really a @xmath6 collider at very high energy , then the determination of photon distribution function inside proton may be an important issue .
there are a few studies about adding the qed corrections to the global parameterizations of pdfs which are based on qcd calculations .
the first one have been done by the mrst group [ and the other analysis are newly released by nnpdf collaboration [ - and ct14qed group [ . here
we present the analytical solutions for dglap evolution equations to determine the parton distribution functions at nlo qcd and lo qed approximations .
our solutions are based on the laplace transform technique which has introduced by block et al [ - .
recently , h. khanpour et al .
[ calculate the proton structure function and parton distribution functions using the laplace transform technique at nlo in qcd without qed corrections .
they consider the initial value of parton distribution functions from kkt12 [ and gjr08 [ codes at @xmath7 .
the laplace transform method has an ability that the analytical solutions for the qed@xmath0qcd parton distribution functions are obtained more strictly by using the related kernels and the calculations can be control well .
following our recent works on analytical solution of dglap evolution equations based on the laplace transform [ - , we have used the same method to solve the qed@xmath0qcd dglap evolution equations .
our initial inputs are selected from apfel ( a pdf evolution library ) code at @xmath7 and the final results were cross - checked with the same results from apfel program and also with newly released ct14qed code to be sure about correctness of our analytical solutions . the paper is organized as follows . in section ,
we review the qed@xmath0qcd coupled dglap evolution equations . in section , we bring out the analytically solutions for the dglap evolution equations to calculate the pdfs inside proton based on the laplace transform .
section is devoted to the results for different kind of the parton distribution functions and also the proton structure function in comparison with other global parameterization results and available experimental data .
finally we give our summary and conclusions in section . in appendix
, we present the results for the different splitting functions in laplace space .
the qed@xmath0qcd dglap evolution equations for the quark , gloun and the photon parton densities can be written as [-: @xmath9 @xmath10 @xmath11 @xmath12 @xmath13 @xmath14 @xmath15 @xmath16 where @xmath17 , @xmath18 , @xmath19 and @xmath5 are the i - th quark , i - th antiquark , the gloun and the photon distribution functions , respectively .
the @xmath0 symbol refers to the convolution integral and the splitting functions in the right - hand side of eq.can be written as , @xmath20 @xmath21 @xmath22 @xmath23 @xmath24 @xmath25 @xmath26 the running strong @xmath27 and electromagnetic @xmath28 couplings determined by the following expressions : @xmath29 @xmath30 where @xmath31 and @xmath32 . for @xmath33 , we get @xmath34 . we suppose @xmath35 then @xmath36 [ .
the lo splitting functions are given by [ @xmath37 @xmath38 @xmath39 @xmath40 @xmath41\ ] ] @xmath42 @xmath43 @xmath44 @xmath45 the @xmath46 , @xmath47 , @xmath48 and @xmath49used in eq .
( ) are the nlo singlet splitting functions , @xmath50and @xmath51are the nlo non - singlet splitting functions that can be found in refs . [ - . for the coupled approach we utilize a pdf basis for the qed@xmath0qcd dglap evolution equations , defined by the following singlet and non - singlet pdf combinations [ , @xmath52 @xmath53 we have found the singlet pdfs evolve as : @xmath54 and the non - singlet pdfs , obey the evolution equations such as : @xmath55 in the equations ( ) and ( ) the new splitting functions are calculated as @xmath56 @xmath57 @xmath58 @xmath59 @xmath60 @xmath61 @xmath62 @xmath63 @xmath64 @xmath65 @xmath66 @xmath67 @xmath68 @xmath69 @xmath70 @xmath71 @xmath72 @xmath73 @xmath74 @xmath75 where @xmath76 and @xmath77 are the number of up- and down - type active quark flavors , respectively , and @xmath78 . in the next section
, we try to solve the above equations with laplace transform method .
now , we are in a position to briefly review the method of extracting the pdfs via the analytical solutions of dglap evolution equations using the laplace transform technique .
block et al . , in ref .
[ , showed that , using the laplace transform , one can solve the dglap evolution equations directly and extract unpolarized parton distribution functions .
we will give the details here and review the method for extracting the unpolarized parton distribution functions of qed@xmath0qcd coupled dglap equations at lo qed and nlo qcd approximations . by introducing the variables @xmath79 and @xmath80 into the coupled dglap equations , one can turn them into coupled convolution equations in @xmath81 and @xmath82 spaces .
we use two laplace transforms from @xmath81 and @xmath82 spaces to s and u spaces , respectively , then the dglap equations can be solved iteratively by a set of convolution integrals which are dependent on unpolarized parton distribution functions at an initial input scale of @xmath83 . using this fact that the laplace transform of a convolution simply is the ordinary product of the laplace transform of the factors , the laplace transform from @xmath81 space to s space convert eq . to ordinary first order differential equations here we intend to extend our calculations to the nlo approximation for the @xmath89 , @xmath90 , gluon and photon sectors of unpolarized parton distributions . in this case , to decouple and to solve dglap evolution equations ( we need an extra laplace transformation from @xmath82 space to u space . in the rest of the calculation , the @xmath91 and @xmath92 are replaced for brevity by @xmath93 and @xmath94 , respectively . therefore the solutions of the first order differential equations in eq . can be converted to , to simplify the nlo calculations we use two excellent approximation relations @xmath96 , where @xmath97 , @xmath98 and @xmath99 and also @xmath100 , where @xmath101 , @xmath102 and @xmath103 for @xmath104 .
@xmath107=\sum\limits _ { j=0}^{1}a_{j}f(s , u+b_{j}),\nonumber \\ l[a^{qed}(\tau){{f}_{j}}(s,\tau);u]=\sum\limits _ { j=0}^{1}\tilde{a}_{j}f(s , u+\tilde{b}_{j}),\label{eq14}\\ b_{0}=0\,and\,\tilde{b}_{0}=0.\nonumber \end{aligned}\ ] ] @xmath109{{\tilde{f}}_{i}}(s , u)-\sum\limits _ { j=2}^{4}\phi_{ij}^ { } { { \tilde{f}}_{j}}(s , u)={{f}_{i0}}(s)\\ + { { \tilde{a}}_{1}}[\sum\limits _ { j = i}^{4}\phi_{ji}^{lo , qed}{{f}_{j}}(s , u+{{\tilde{b}}_{1}})]\\ + { { a}_{1}}[\sum\limits _ { j = i}^{4}\phi_{ji}^{nlo , qcd}{{f}_{j}}(s , u+{{b}_{1}})]\qquad i , j=1,\ldots,4\label{eq16}\end{gathered}\ ] ] the complete solutions of eq . (
can be obtained via iteration processes .
the iteration can be continued to any required order but we will restrict our selves in which we get to a sufficient convergence of the solutions .
our results show that the second order of iteration are sufficient to get a reasonable convergence . using the first inverse laplace transform technique from u space to @xmath82 space
, we can obtain the following expression for the distributions , with the initial input functions for @xmath90 , @xmath89 , gloun and photon sectors of distributions , which are denoted by @xmath111 , @xmath112 , @xmath113 and @xmath114 , respectively . by the second inverse laplace
transform from s space to @xmath79 space , we get parton distribution functions in the usual x space .
we perform here the non - singlet solutions of the qed@xmath0qcd dglap evolution equation , ( ) , using the laplace transform technique at lo qed and nlo qcd approximations . for the non - singlet distributions @xmath115 , after changing to the variable @xmath116 and the variable @xmath82 , we can schematically write the equation ( as the @xmath127 parameter is related to the leading order qed running coupling constant . the non - singlet solutions , @xmath131 , can be obtained using the non - singlet kernel @xmath132 $ ] in the convolution integral @xmath133 finally , with these two laplace transforms , the evolution equations ( can be solved iteratively by a set of convolution integrals which are related to the quark distributions at an initial input scale of @xmath83 in @xmath134 space . in this section , we will present our results that we are obtained for the parton distribution functions and proton structure function , @xmath1 , using the laplace transform technique .
the results are displayed in figs .
it should be noted that we need some initial inputs for pdfs , equations ( and ( ) .
we borrowed them from apfel code [ at @xmath7 , to be sure about the correctness of our solutions .
if the solutions are correct then we expect that our pdfs set and proton structure function have good agreement with those from all global parameterizations ( as well as apfel ) and experimental data .
the valance quark distributions , @xmath135 and @xmath136 , at lo qed and nlo qcd approximations are depicted in figs . and .
we also compare them with apfel model results for the different values of @xmath2 .
the solid curves show our results for the valence quark distributions , and the scatter curves present the apfel model results .
the agreement with both the d and u valance quark distributions , over the large range of x and @xmath2 , is excellent .
the results show that our analytical solutions for the qed@xmath8qcd dglap evolution equations are correct and these solutions are correctly used to calculate the parton distribution functions . in figure
the different kinds of parton distribution functions are shown at @xmath137 in comparison with those from apfel model .
it is found that the sea quark distribution functions in comparison with the photon distribution function in the large values of x with increasing the value of @xmath2 , contribution of photon is most signicant . it may also be noted that in range of high x the photon distribution function is larger than the bottom quark distribution function as increasing the value of @xmath2 . in figure we displayed the pdfs at two different scales of @xmath2 :
@xmath138 ( dashed - curves ) and @xmath137 ( solid - curves ) .
it is shown that with increasing the value of @xmath2 , the contribution of valence quarks are decreased .
therefore , we can conclude the photon contribution is now significantly considerable.it is observed from this figure with increasing the value of @xmath2 that the parton distribution functions decrease for the large values of x and increase for the small values of x. fig .
displays our analytical parton distribution functions , @xmath139 , at @xmath137 .
we compared our results with the newly released pdfs global parameterizations from ct14qed [ .
ct14 qed , is the first set of ct14 parton distribution functions obtained by including qed evolution at leading order with next - to - leading order qcd evolution in their global analysis .
the comparison photon distribution function , @xmath140 , with apfel and ct14qed models at @xmath137 for @xmath141 is well demonstrated in fig .. this plot indicates that our results are in good agreement with apfel model .
also it is clear from this figure that our results in comparison with the ct14qed photon distribution function are very similar at large value of x and are different for small value of x. we also investigate the effect of an increasing in value of @xmath142 on the photon distribution functions and conclude that the ct14qed photon distribution function becomes large , whereas our results is distinctly dierent and much smaller at small values of x ( corresponding plot omitted for briefly ) .
we now proceed by calculating proton structure function .
the laplace transform technique is also applied to the proton structure function , @xmath1 , which leads to an analytical solution for this function .
the method illustrated in this analysis enable us to achieve strictly analytical solution for proton structure function in terms of x variable .
figure depicts the comparison the proton structure function with the corresponding available experimental data from the h1 and zeus collaborations in the several values of @xmath2 .
the results demonstrate that there are good agreement between them .
it is clear that the proton structure function increases with an increase in value of @xmath2 for small values of x and decrease for large values of x. all figures indicate that the analytical solutions work well beyond the charm quark mass threshold , @xmath143 . in this paper , we utilized the laplace transform technique to calculate the laplace transformation of splitting functions and extract the parton distribution functions of quark , antiquark , gloun and photon inside the proton .
our calculations are done in lo qed and nlo qcd approximations .
we finally extracted the unpolarized proton structure functions at the different values of @xmath2 .
our results are compared with apfel and newly released ct14qed codes and also with experimental data which indicate good agreements between them . to determine the proton structure function at any arbitrary @xmath2 scale , we only need to know the initial distributions for singlet , gluon , non - singlet and photon distributions at the input scale of @xmath83 .
we borrowed the initial inputs from apfel code at @xmath7 to be sure about the correctness of our solutions .
the solutions are seem to be correct because the parton distribution functions and the proton structure function have good agreement with those from all global parameterizations ( as well as apfel ) and experimental data.in the future work with a global parameterization we can determine these initial inputs .
these pdfs can specifically design for use in precision cross section predictions and uncertainties at the lhc .
we present here the results for the laplace transforms of the splitting functions at lo qed and nlo qcd approximations which we used in eq . ) .
we fixed the usual quadratic casimir operators to their exact values , using @xmath144 , @xmath145 ( where the@xmath146 is the number of active quark flavours ) , @xmath147 ( is number of colors ) and@xmath148 .
the @xmath149 is defined by @xmath150 and @xmath151 = 0.577216 is the euler lagrange constant . | we analytically solved the qed@xmath0qcd coupled dglap evolution equations at leading order ( lo ) quantum electrodynamics ( qed ) and next to leading order ( nlo ) quantum chromodynamics ( qcd ) approximations , using the laplace transform method and then computed the proton structure function in terms of the unpolarized parton distributions functions .
our analyitical solutions for parton densities are in good agreement with those from apfel ( a pdf evolution library ) ( computer physics communications 185 , 1647 - 1668 ( 2014 ) ) and ct14qed ( phys .
rev .
d 93 , 114015 ( 2016 ) ) global parameterizations .
we also compared the proton structure function , @xmath1 , with experimental data released by the zeus and h1 collaborations at hera .
there is a nice agreement between them in the range of low and high x and @xmath2 . |
there are more than 150 different objects enumerated by catalan numbers ; @xcite contains an extensive list of such combinatorial objects and their properties .
two of the most carefully studied ones are noncrossing partitions and permutations avoiding a 3-letter pattern .
a partition @xmath4 of the set @xmath5 \colon = \{1,2,\cdots , n\}$ ] , having blocks @xmath6 , is called _ noncrossing _ if there are no four elements @xmath7 so that @xmath8 and @xmath9 for some distinct blocks @xmath10 and @xmath11 . the set of noncrossing partitions of @xmath5 $ ] constitutes a lattice under the refinement order ( where @xmath12 if each block of @xmath13 is a union of blocks of @xmath4 ) .
an investigation of structural and enumerative properties of this lattice was initiated by kreweras @xcite , and continued by several authors , e.g. , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite .
we denote the lattice of noncrossing partitions of @xmath5 $ ] as @xmath14 , since it is a subposet ( indeed , a sub - meet - semilattice ) of the intersection lattice associated with the type - a hyperplane arrangement in @xmath15 ( which consists of the hyperplanes with equations @xmath16 , for @xmath17 ) . for our purposes , recall from @xcite that the poset @xmath14 is ranked with rank function @xmath18 ( where @xmath19 denotes the number of blocks of the partition @xmath4 ) , rank - symmetric and rank - unimodal with rank sizes given by the narayana numbers @xmath20 .
furthermore , it is self - dual ( see @xcite , @xcite ) and has the strong sperner property ( see @xcite ; that is , for every @xmath21 , the maximum cardinality of the union of @xmath21 antichains is the sum of the @xmath21 largest rank - sizes ) . a permutation @xmath22 of @xmath5 $ ] , or , in what follows , an @xmath23-permutation , is called _ 132-avoiding _ if there are no three positions @xmath24 so that @xmath25 .
classes of restricted permutations avoiding other patterns are defined similarly .
such classes of permutations arise naturally in theoretical computer science in connection with sorting problems ( e.g. , @xcite , @xcite ) , as well as in the context of combinatorics related to geometry ( e.g. , the theory of kazhdan - lusztig polynomials @xcite and schubert varieties @xcite ) .
the investigation of classes of pattern - avoiding permutations from an enumerative and algorithmic point of view includes @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , @xcite , to name a few . in section [ sec -
a ] of this paper we introduce the partially ordered set @xmath0 whose elements are the 132-avoiding @xmath23-permutations , ordered by @xmath26 if @xmath27 , where @xmath28 denotes the descent set of a permutation .
one can think of @xmath0 as a boolean algebra of rank @xmath29 in which each element @xmath30 is replicated as many times as there are 132-avoiding permutations with @xmath30 as the descent set .
we show that this poset of restricted permutations is an extension of the lattice of noncrossing partitions @xmath14 by exhibiting a natural order - preserving bijection from the dual order @xmath31 to the poset @xmath0 .
this yields the fact that @xmath0 has the same rank - generating function as @xmath14 ( implicit in @xcite , where the joint distribution of the descent and major index statistics on 132-avoiding permutations is shown to agree with the joint distribution of the block and rb statistic on noncrossing partitions ) .
it then follows that @xmath0 is rank - unimodal , rank - symmetric and strongly sperner .
we also prove that @xmath0 is itself a self - dual poset .
we also present type - b analogues of these results .
these constitute section [ sec - b ] of the paper .
the notion of a type - b noncrossing partition of @xmath5 $ ] is that first considered by montenegro @xcite , systematically studied by reiner @xcite , and further investigated by hersh @xcite .
these authors show that the type - b noncrossing partitions of @xmath5 $ ] form a lattice , @xmath32 , which shares naturally a variety of properties of @xmath14 . in particular
, @xmath32 is a rank - unimodal , self - dual , strongly sperner poset .
we define a poset @xmath1 into which @xmath32 can be embedded via an order - preserving bijection , with properties analogous to those obtained for type - a .
the parallel between the type - a and type - b cases includes the fact that the poset @xmath1 is defined in terms of pattern - avoiding elements of the hyperoctahedral group ( or signed permutations ) , ordered by containment of the descent set .
the relevant pattern restriction is the simultaneous avoidance of the patterns @xmath33 and @xmath34 .
this class of restricted signed permutations was considered in @xcite , where b - analogues are proposed for type - a results in @xcite concerning combinatorial statistics for noncrossing partitions and restricted permutations . in brief , a class of partitions and one of permutations are equinumerous , and further , the count of the partitions by number of blocks agrees with the count of permutations by number of descents .
a similar situation arises for certain type - b analogues of these objects .
our results show that these enumerative relations are manifestations of structural relations between partial orders which can be defined naturally on the objects under consideration .
we also discuss posets @xmath2 and @xmath3 of restricted permutations and signed permutations ordered by containment of their sets of excedences .
the final section of the paper consists of remarks and problems for further investigation .
it is not difficult to find a bijection from the set of noncrossing partitions of @xmath5 $ ] onto that of 132-avoiding @xmath23-permutations . here
we exhibit and analyze the structure of such a bijection , @xmath35 , which will serve as the main tool in proving the results of this section . to avoid confusion , integers belonging to a partition
will be called _ elements _ , while integers belonging to a permutation will be called _
entries_. an @xmath23-permutation will always be written in the one - line notation , @xmath36 , with @xmath37 denoting its @xmath38th entry .
let @xmath39 .
we construct the 132-avoiding permutation @xmath40 corresponding to it as follows .
let @xmath21 be the largest element of @xmath4 which is in the same block of @xmath4 as 1 .
put the entry @xmath23 of @xmath41 in the @xmath21th position , i.e. , set @xmath42 . as @xmath41 is to be 132-avoiding
, this implies that the entries larger than @xmath43 are on the left of @xmath23 in @xmath41 , and the entries smaller than or equal to @xmath43 are on the right of @xmath23 .
delete @xmath21 from @xmath4 and apply this procedure recursively , with obvious minor adjustments , to the restrictions of @xmath4 to the sets @xmath44 and @xmath45 , which are also noncrossing partitions .
namely , if @xmath46 is the largest element in the same block as @xmath47 , we set @xmath48 , so that the restriction @xmath49 of @xmath4 to @xmath50 yields a 132-avoiding permutation of @xmath51 placed on the right of @xmath23 in @xmath52 . similarly , if in the restriction @xmath53 of @xmath4 to the set @xmath54 the largest element in the same block as 1 is equal to @xmath46 , we set @xmath55 .
thus , recursively , @xmath53 yields a 132-avoiding permutation which we realize on the set @xmath56 and we place it to the left of @xmath23 in @xmath52 . in other words , with a slight abuse of notation , @xmath57 is the concatenation of @xmath58 , @xmath23 , and @xmath59 , where @xmath58 permutes the set @xmath60 and @xmath59 permutes the set @xmath61 $ ] . to see that this is a bijection note that we can recover the maximum of the block containing the element 1 from the position of the entry @xmath23 in @xmath41 , and then proceed recursively . _
if @xmath62 , then @xmath63 .
_ _ if @xmath64 , then @xmath65 . _ _ if @xmath66 , then @xmath67 . _
the following definition is widely used in the literature .
let @xmath36 be an @xmath23-permutation .
we say that @xmath68 $ ] is a _ descent _ of @xmath41 if @xmath69 .
the set of all descents of @xmath41 is called the _ descent set _ of @xmath41 and is denoted @xmath70 .
now we are in a position to define the poset @xmath0 of 132-avoiding permutations we want to study .
let @xmath41 and @xmath71 be two 132-avoiding @xmath23-permutations .
we say that @xmath72 in @xmath0 if @xmath73 . clearly , @xmath0 is a poset as inclusion is transitive .
the hasse diagram of @xmath74 is shown in figure 1 .
[ obs - lrmina ] in a 132-avoiding permutation , @xmath38 is a descent if and only if @xmath75 is smaller than every entry on its left .
such an element is called a _ left - to - right minimum_. so @xmath72 in @xmath0 if and only if the set of positions of left - to - right minima in @xmath41 is a proper subset of the set of positions of left - to - right minima in @xmath71 .
the following proposition describes the relation between the blocks of @xmath39 and the descent set of the 132-avoiding permutation @xmath57 .
[ prop - block ] the bijection @xmath35 has the following property : let @xmath76
. then @xmath77 if and only if @xmath78 is the smallest element of its block in @xmath39 . for @xmath79 and @xmath80
the statement is clearly true and we use induction on @xmath23 .
suppose we know the statement for all positive integers smaller than @xmath23 .
then we distinguish two cases : 1 .
if 1 and @xmath23 are in the same block of @xmath4 , then the construction of @xmath57 starts by putting the entry @xmath23 in the last slot of @xmath57 , then deleting the element @xmath23 from @xmath4 .
this does not alter either the set of minimum elements of the blocks nor the set of descents .
therefore , this case reduces to the general case for @xmath81 , and is settled by the inductive hypothesis .
if the largest element @xmath21 of the block containing 1 is smaller than @xmath23 , then as we have seen above , @xmath57 is the concatenation of @xmath82 , and @xmath59 is not empty . clearly , by the definition of @xmath57 , @xmath83 and the element @xmath47 is the minimum of its block . from this and the inductive hypothesis applied to @xmath59 and @xmath58
, the proof follows .
proposition [ prop - block ] implies that @xmath0 is isomorphic to the order on noncrossing partitions in which @xmath84 if the set of minima of the blocks of @xmath85 is contained in the set of minima of the blocks of @xmath4 .
this yields the first result of this section .
[ thm - embpa ] the lattice of noncrossing partitions @xmath14 is a subposet of @xmath0 . we show that our bijection @xmath35 is an order - reversing map @xmath86 .
the conclusion then follows from the self - duality of the lattice of noncrossing partitions .
suppose @xmath87 in @xmath14 .
this means @xmath4 is a finer partition than @xmath88 , so every element which is the minimum of its block in @xmath88 is also the minimum of its block in @xmath4 . by proposition [ prop - block
] this implies @xmath89 , so @xmath90 in @xmath0 . clearly , @xmath0 is a ranked poset ( with rank function @xmath91 ) , and we have @xmath92 .
[ cor - syma ] the poset @xmath0 is rank - symmetric , rank - unimodal and strongly sperner , and its rank generating function is equal to that of @xmath14 .
the properties of the rank sizes of @xmath0 are immediate consequences of proposition [ prop - block ] and the corresponding properties known to hold for @xmath14 .
moreover , every antichain of @xmath0 is , via the bijection @xmath35 , an antichain of @xmath14 , and the strong sperner property of @xmath0 follows from the strong sperner property of @xmath14 .
we now turn to showing that @xmath0 is self - dual , based on the next lemma . for @xmath93 $ ] , let @xmath94 denote the number of 132-avoiding @xmath23-permutations with descent set @xmath30 .
[ lemma - rev - compl ] let @xmath30 be any subset of @xmath95 $ ] and let @xmath96 denote its `` reverse complement , '' that is , @xmath97 . then @xmath98 .
we use induction on @xmath23 . for @xmath99
the statement is true . now
suppose we know it for all positive integers smaller than @xmath23 .
denote by @xmath100 the smallest element of @xmath30 , and let @xmath41 be a 132-avoiding @xmath23-permutation whose descent set is @xmath30 . 1 .
suppose that @xmath101 .
then we have @xmath102 and , because @xmath41 avoids the pattern 132 , the values of @xmath103 are _ consecutive _ integers .
so , for given values of @xmath104 and @xmath100 , we have only one choice for @xmath105 .
this implies @xmath106 where @xmath107 is the set obtained from @xmath30 by subtracting @xmath108 from each of its elements .
+ on the other hand , we have @xmath109 , meaning that in any permutation @xmath71 counted by @xmath110 the chain of inequalities @xmath111 holds . to avoid forming a 132-pattern in @xmath71 , we must have @xmath112 .
therefore , @xmath113 where @xmath114 denotes the set obtained from @xmath115 by removing its last @xmath108 elements . clearly
, @xmath116 by the induction hypothesis , so equations ( [ 1a ] ) and ( [ 1b ] ) imply @xmath98 .
2 . if @xmath117 , but @xmath118 $ ] , then let @xmath119 be the smallest index which is _ not _ in @xmath30 . then again , to avoid forming a 132-pattern , the value of @xmath120 must be the smallest positive integer @xmath121 which is larger than @xmath122 and is not equal to any @xmath123 for @xmath124 .
so again , we have only one choice for @xmath120 . on the other hand , the largest index in @xmath115 will be @xmath125 .
therefore , in permutations @xmath71 counted by @xmath110 , we must have @xmath126 as observation 1 implies that @xmath127 must be the rightmost left - to - right minimum in such permutations , and that is always the entry 1 .
+ in order to use this information to reduce our permutations in size , we define @xmath128 $ ] as follows : @xmath129 if and only if either @xmath130 and then , by the definition of @xmath119 , @xmath131 , or @xmath132 and @xmath133 .
in other words , we decrease elements larger than @xmath119 by 1 ; intuitively , we remove @xmath119 from @xmath95 $ ] , and translate the interval on its right one notch to the left . if we now take @xmath134 , that will consist of entries @xmath46 so that @xmath135 and @xmath136 .
so in other words , we simply remove @xmath125 from @xmath95 $ ] ( there has been nothing on the right of @xmath125 in @xmath115 to translate ) .
note that the size of @xmath115 decreases with this operation as @xmath137 . as we have seen in the previous paragraph , we had only one choice for @xmath120 and @xmath138 , so removing them this way does not change the number of permutations with a given descent set .
thus we have @xmath139 , and also @xmath140 . by induction hypothesis ,
the right hand sides of thes two equations agree , and therefore the left hand sides must agree , too .
+ _ if @xmath141 and @xmath142 , and so @xmath143 , then @xmath144 , @xmath145 , and indeed , @xmath146 and @xmath147 . _
3 . finally ,
if @xmath148 $ ] , then the statement is trivially true as @xmath149 so we have seen that @xmath98 in all cases .
it is now easy to verify that the reverse complementation of the descent set can be used to construct an anti - automorphism of @xmath0 .
the poset @xmath0 is self - dual .
it is clear that , in @xmath0 , permutations which have the same descent set will cover the same elements and they will be covered by the same elements . the permutations with a prescribed descent set @xmath30 form an orbit of @xmath150 and they can be permuted among themselves arbitrarily by elements of @xmath150 .
on the other hand , lemma [ lemma - rev - compl ] shows that the orbits corresponding to @xmath93 $ ] and to its reverse - complement @xmath115 are equinumerous .
hence , a map @xmath151 which establishes a bijection between @xmath152 and @xmath153 for each @xmath154 $ ] provides an order - reversing bijection on @xmath0 .
it is shown in @xcite that the joint distribution of the excedence and denert statistics on 321-avoiding permutations agrees with the joint distribution of the block and @xmath155 statistics on noncrossing partitions .
this suggests the definition of the poset @xmath2 consisting of the 321-avoiding @xmath23-permutations ordered by containment of the set of excedences , and invites the question of how @xmath2 compares with the poset @xmath0 .
a permutation @xmath156 has an _ excedence _ at @xmath38 if @xmath157 .
for example , the _ excedence set _ of @xmath158 is @xmath159 .
let @xmath160 denote the number of excedences of @xmath156 .
following @xcite , there is a bijection @xmath161 from @xmath162 to 321-avoiding @xmath23-permutations such that @xmath163 .
namely , if the set of minima of the blocks of @xmath39 , omitting the block containing 1 , is @xmath164 and the set of maxima of the blocks , again , omitting the block containing 1 , is @xmath165 , then let @xmath166 be the permutation whose value at @xmath167 is @xmath168 for @xmath169 , and whose other values constitute an increasing subsequence in the remaining positions .
for instance , if @xmath170 , then we have @xmath171 and @xmath172 , and we obtain @xmath173 . recall from @xcite that the set of excedences of @xmath166 is precisely @xmath174 .
similarly to the case of descents discussed for 132-avoiding permutations , a covering relation @xmath175 in @xmath14 corresponds to the deletion of an excedence : @xmath176 , for a suitable @xmath177 .
hence , taking advantage of the self - duality of @xmath14 , one can establish directly that the poset @xmath2 enjoys the same properties as @xmath0 : there is an embedding of @xmath14 into the poset @xmath2 of 321-avoiding @xmath23-permutations ordered by containment of the set of excedences ; the embedding is rank - preserving and @xmath2 is a strongly sperner poset .
the fact that the posets @xmath0 and @xmath2 have strongly similar properties is not accidental .
[ prop - isoa ] the posets @xmath0 and @xmath2 are isomorphic . for each @xmath93 $ ] , let @xmath178 be the set of 321-avoiding @xmath23-permutations with excedence set @xmath179 $ ] .
let also @xmath180 be the set of 132-avoiding @xmath23-permutations with descent set equal to @xmath115 , the reverse - complement of @xmath30 .
thus , in the notation of the previous subsection , the cardinality of @xmath180 is @xmath110 .
we construct a bijection @xmath181 ( illustrated by example [ eg - sbij ] ) . if @xmath182 , then , as seen earlier in the definition of @xmath161 , the entries @xmath183 with @xmath184 form an increasing subsequence .
this , and the definition of excedence imply that @xmath183 is a _
right - to - left minimum _ ( that is , smaller than all entries on its right ) if and only if @xmath185 .
now let @xmath186 be the reverse of @xmath41 .
then @xmath187 is a 123-avoiding permutation having a left - to - right minimum at position @xmath188 exactly if @xmath189 .
there is exactly one 132-avoiding permutation @xmath190 which has this same set of left - to - right minima at these same positions @xcite .
namely , @xmath190 is obtained by keeping the left - to - right minima of @xmath187 fixed , and successively placing in the remaining positions , from left to right , the smallest available element which does not alter the left - to - right minima .
we set @xmath191 .
observation 1 then tells us that @xmath192 if and only if @xmath193 , in other words , when @xmath194 , and so @xmath190 belongs indeed to @xmath180 .
it is easy to see that @xmath195 is invertible .
clearly , @xmath187 can be recovered from @xmath190 as the only 123-avoiding permutation with the same values and positions of its left - to - right minima as @xmath190 .
( all entries which are not left - to - right minima are to be written in decreasing order )
. then @xmath41 can be recovered as the reverse of @xmath187 .
the bijections @xmath181 for all the choices of @xmath93 $ ] produce an order - reversing bijection from @xmath2 to @xmath0 .
but @xmath0 is self - dual , so the proof is complete .
[ eg - sbij ] _ take @xmath196 for @xmath197
. then its reversal @xmath198 has left - to - right minima 8 , 7 , 5 , 3 , 1 in positions 1 , 2 , 5 , 6 , 8 .
we obtain @xmath199 , a permutation in @xmath200 . _
the hyperplane arrangement of the root system of type @xmath201 consists of the hyperplanes with equations @xmath202 for @xmath17 and the coordinate hyperplanes @xmath203 , for @xmath204 .
the subspaces of @xmath15 arising as intersections of hyperplanes from among these can be encoded by partitions of @xmath205 satisfying the following properties : i ) if @xmath206 is a block , then @xmath207 is also a block , where the bar operation is an involution ; and ii ) there is at most one block , called _ the zero - block _ , which is invariant under the bar operation .
the collection of such partitions are the _ type - b partitions of @xmath5$]_. if @xmath208 are placed around a circle , clockwise in this order , and if cyclically successive elements of the same block are joined by chords drawn inside the circle , then , following @xcite , the class of _ type - b noncrossing partitions _
, denoted @xmath32 , is the class of type - b partitions of @xmath5 $ ] which admit a circular diagram with no crossing chords .
alternatively , a type - b partition is noncrossing if there are no four elements @xmath209 in clockwise order around the circle , so that @xmath210 lie in one block and @xmath211 lie in another block of the partition .
the total number of type - b noncrossing partitions of @xmath5 $ ] is @xmath212 ( see @xcite ) .
as in the case of type a , the refinement order on type - b partitions yields a geometric lattice ( in fact , isomorphic to a dowling lattice with an order-2 group ) , and the noncrossing partitions constitute a sub - meet - semilattice as well as a lattice in its own right . as a poset under the refinement order ,
@xmath32 is ranked , with @xmath213 of pairs of non - zero blocks @xmath214 .
for example , @xmath215 is an element of @xmath216 having 2 pairs of non - zero blocks and its rank is equal to 3 .
the rank - sizes in @xmath32 are given by @xmath217 ( see @xcite ) .
the numerous properties of @xmath14 which also hold for @xmath32 ( as shown in @xcite , @xcite ) , establish the latter as a natural b - analogue .
in particular , @xmath32 is a self - dual , rank - unimodal , strongly sperner poset , analogously to the properties of @xmath14 of concern in section [ sec - a ] .
we now turn to a type - b counterpart of the restricted permutations considered in the preceding section .
we will view the elements of the hyperoctahedral group @xmath201 as signed permutations written as words of the form @xmath218 in which each of the symbols @xmath219 appears , and may or may not be barred .
thus , the cardinality of @xmath201 is @xmath220 . to find a b - analogue of the poset @xmath0
, we need a subset of @xmath201 whose cardinality is @xmath221 , which is characterized via pattern - avoidance , and over which the distribution of the descent statistic agrees with the distribution across ranks of the type - b noncrossing partitions of @xmath5 $ ] .
such a class of signed permutations is @xmath222 which appears in @xcite .
we include its description for the reader s convenience .
consider the elements of @xmath201 which avoid simultaneously the patterns @xmath33 and @xmath34 .
that is , the set of elements @xmath223 such that there are no indices @xmath17 for which i ) either both @xmath224 are barred , or neither is barred , and ii ) @xmath225 ( the absolute value of a symbol means @xmath226 if @xmath121 is not barred , and @xmath227 if @xmath121 is barred ; effectively , the absolute value removes the bar from a barred symbol ) .
the following is immediate : a @xmath228-avoiding permutation in @xmath201 is a shuffle of an increasingly ordered subset @xmath229 of @xmath5 $ ] whose elements we then bar , with its increasingly ordered complement in @xmath5 $ ] . for example
, @xmath230 is one of @xmath231 elements of @xmath232 associated with the subset @xmath233 $ ] .
obviously , summing over the choices of @xmath229 of cardinality ranging from zero to @xmath23 and over the shuffles , it follows that @xmath234 as desired .
furthermore , the distribution of descents over @xmath235 is as desired .
we say that @xmath236 has a _ descent _ at @xmath38 , for @xmath237 , if @xmath238 with respect to the total ordering @xmath239 , and that it has a descent at @xmath23 if @xmath240 is barred . as usual , the _ descent set _ of @xmath241 , denoted @xmath242 , is the set of all @xmath243 $ ] such that @xmath241 has a descent at @xmath38 .
for example , for @xmath244 we have @xmath245 .
it is then transparent that if @xmath246 , then its descent set is precisely the set of positions occupied by barred symbols . in conclusion , [ obs - bij - b2121 ] for an element @xmath241 of the hyperoctahedral group @xmath201 , let @xmath247 denote the set of symbols which are barred in @xmath241 , and @xmath242 denote the descent set of @xmath241 .
then the map @xmath248 gives a bijection between the class of restricted signed permutations @xmath249 and ordered pairs of subsets of @xmath5 $ ] of equal cardinality . as the b - analogue of the poset of 132-avoiding permutations @xmath0 of the preceding section , we consider the poset @xmath1 consisting of the @xmath250-avoiding elements of the hyperoctahedral group @xmath201 , with the order relation given by @xmath251 if and only if @xmath252 .
based on the preceding discussion and an encoding of type - b noncrossing partitions appearing in @xcite , one readily obtains the properties of @xmath1 which parallel those of @xmath0 .
the poset @xmath1 of @xmath253-avoiding elements of the hyperoctahedral group @xmath201 , ordered by containment of the descent set , is an extension of the refinement order on the type - b noncrossing partition lattice @xmath32 .
the poset @xmath1 has the same rank - generating - function as @xmath32 , therefore it is rank - symmetric and rank - unimodal , and it is a self - dual and strongly sperner poset .
it is immediate from its definition and observation [ obs - bij - b2121 ] that @xmath1 is a ranked poset ( namely , @xmath254 ) and has rank - sizes given by @xmath255 , equal to the rank - sizes in @xmath32 .
also , @xmath1 is a self - dual poset : clearly , if @xmath256 is the @xmath250-avoiding signed permutation which corresponds to the pair @xmath257 - l(b ) , [ n ] - { \rm des}(b ) ) $ ] , then the mapping @xmath258 is an order - reversing involution on @xmath1 . toward checking
that there is an order - preserving bijection from @xmath32 to @xmath1 , we first recall a fact from @xcite : every partition @xmath259 can be encoded by a pair @xmath260 of subsets of @xmath5 $ ] whose cardinality is the number of pairs of non - zero blocks of @xmath4 . informally , these sets consist of the left and right delimiters of non - zero blocks when the elements are read in clockwise order ( in the circular diagram of @xmath4 ) .
more precisely , if @xmath261 or if @xmath4 has only a zero - block , we set @xmath262 . otherwise , @xmath259 has some non - zero block consisting of cyclically consecutive elements in its diagram .
if such a block consists of @xmath263 in clockwise order , then @xmath264 belongs to @xmath265 and @xmath266 belongs to @xmath267 . by deleting this block and its image under barring , a type - b noncrossing partition of @xmath61 $ ]
is obtained and the construction of the sets @xmath265 and @xmath267 is completed by repeating this process as long as non - zero blocks arise . for instance , if @xmath268 , then @xmath269 and @xmath270 .
now suppose that @xmath175 in @xmath32 , and that this is a covering relation ( i.e. , @xmath271 )
. then there exist @xmath272 and @xmath273 such that @xmath274 and @xmath275 , as a result of the merging of blocks entailed by the covering relation .
thus it is clear that if @xmath259 is mapped to the signed permutation @xmath276 with the property that @xmath277 , then one obtains an order - reversing embedding of @xmath32 into @xmath1 . combining this with the self - duality of @xmath1 we obtain the desired embedding of @xmath32 into @xmath1 .
finally , the strong sperner property of @xmath1 follows as in type a , from the strong sperner property of @xmath32 ( see @xcite ) and the rank - preserving embedding of @xmath32 into @xmath1 . as in the type - a case , there is a self - dual poset of @xmath278 restricted signed permutations ordered by containment of the set of excedences .
in fact , there is more than one definition of the excedence statistic in the literature , in the case of the hyperoctahedral group .
we briefly mention two possibilities considered in @xcite . given a signed permutation @xmath241 , let @xmath21 be the number of symbols which are _ not _ barred in @xmath241 .
we associate to @xmath241 an @xmath279-permutation @xmath280 by setting @xmath281 and , for @xmath204 , letting @xmath282 if @xmath283 is the @xmath46th smallest among the symbols @xmath284 with respect to the linear ordering @xmath285 .
for example , if @xmath286 , then @xmath287 .
now , the excedence set of @xmath241 is defined to be that of @xmath280 .
it turns out @xcite that for @xmath288 this definition makes the excedence set coincide with the descent set for each @xmath241 .
therefore , this leads to the poset @xmath1 again .
an alternative definition for excedences of `` indexed permutations '' appears in @xcite .
specialized to the hyperoctahedral group it is the following .
[ def - bexc2 ] if @xmath289 , its _ excedence set _ is the union of the sets @xmath290 and @xmath291 , where @xmath290 is the set of excedences computed in the symmetric group for the permutation @xmath292 obtained by removing all bars from the symbols in @xmath241 , and @xmath293 , the set of barred fixed points of @xmath241 .
thus , for @xmath294 we obtain excedences at @xmath295 by either of the two definitions . but
@xmath296 has excedences at @xmath297 by the first definition ( based on @xmath298 ) , and @xmath299 if the second definition is adopted . for the remainder of this section
, we work with the notion of excedence as in definition [ def - bexc2 ] .
[ prop - qbn - sdual ]
let @xmath3 denote the poset of @xmath250-avoiding signed permutations in @xmath201 , ordered by containment of their excedence set .
the poset @xmath3 is self - dual .
let @xmath300 and @xmath256 be the reverse of @xmath241 .
let @xmath301 be the `` barred complement '' of @xmath256 , that is , @xmath302 , and @xmath303 is barred if and only if @xmath304 is not barred .
then it is straightforward to verify that @xmath305 if and only if @xmath306 .
therefore , the reverse complement operation reverses the inclusion of excedence sets for signed permutations .
( thus , the entire hyperoctahedral group @xmath201 ordered by containment of the excedence set is a self - dual poset . ) but , clearly , this involution preserves the @xmath307-avoidance property , and thus @xmath3 is self - dual . by @xcite , the rank generating function of @xmath3 is equal to that of @xmath1 .
therefore it is natural to ask whether the posets @xmath1 and @xmath3 are isomorphic , just as their type - a counterparts are ( proposition [ prop - isoa ] ) .
the answer in this case is negative . indeed , if @xmath308 it is straightforward to verify that all atoms of @xmath309 are covered by six elements , while the atom @xmath310 of @xmath311 is covered by seven elements ( namely , @xmath312 , and @xmath313 ) .
_ is @xmath32 a subposet of @xmath3 ? _ we do not know whether the lattice of type - b noncrossing partitions can be embedded in the poset @xmath3 of @xmath250-avoiding signed permutations ordered by their excedence set of definition [ def - bexc2 ] .
_ self - duality of @xmath14 and @xmath32 extending to self - duality for @xmath0 and @xmath1 .
_ + we have seen that _ each _ of the posets @xmath14 and @xmath0 is self - dual and that @xmath14 is a subposet of @xmath0 .
the same is true for the pair @xmath32 , @xmath1 . both for type
a and for type b one can exhibit an order - reversing involution on the larger poset which restricts to an order - reversing involution on the smaller one .
+ we first construct such an involution @xmath314 for @xmath14 which will be similar , though not identical , to the involution defined in @xcite .
+ write the elements @xmath315 clockwise around a circle , and write elements @xmath316 interlaced in counterclockwise order , so that @xmath317 is between @xmath318 and @xmath23 , @xmath319 is between @xmath23 and @xmath81 , and so on , @xmath320 is between @xmath321 and @xmath322 . for @xmath39 ,
join by chords as usual cyclically successive ( unprimed ) elements belonging to the same block of @xmath4 . then define @xmath323 to be the coarsest noncrossing partition on the elements @xmath316 so that the chords joining primed elements of the same block do not intersect the chords of @xmath4 .
see figure 2 for an example . + the map @xmath314 is certainly a bijection , and it is order - reversing in @xmath14 since merging two blocks of @xmath4 subdivides a block of @xmath323 .
we claim that @xmath314 is also order - reversing on @xmath0 .
to see this , observe that for any @xmath324 , the element @xmath38 is the smallest in its block in @xmath4 if and only if the element @xmath325 is _ not _ the smallest in its block in @xmath323 .
indeed , the definition of @xmath314 implies that exactly one of @xmath38 and @xmath325 can be connected to smaller elements by a chord .
therefore , @xmath314 takes the set of block - minima ( not equal to 1 ) of @xmath4 into its reverse complement in @xmath326 $ ] , so @xmath314 is indeed order - reversing on @xmath0 .
+ in the type-@xmath327 case , one can obtain an analogous bijection @xmath328 in a similar way : take the circular ( clockwise ) representation of @xmath4 , then write the elements @xmath329 so that the primed numbers interlace the unprimed , placing @xmath317 between @xmath318 and @xmath330 and continuing counterclockwise . for @xmath259 , define @xmath331 as above , that is , as the unique coarsest partition on the primed set whose chords do not cross those of @xmath4
. then @xmath328 is certainly an order - reversing bijection of @xmath32 , and as above , it reverses the containment of the sets @xmath265 and @xmath267 , so it does extend to an order - reversing bijection of @xmath1 .
the mbius function and order complexes of @xmath0 and @xmath1 .
_ + it is easy to write an expression for the number @xmath332 of chains @xmath333 of length @xmath334 in @xmath1 , for @xmath335 . of course ,
@xmath336 , and @xmath337 since under the correspondence @xmath338 a chain in @xmath1 corresponds to an @xmath339-tuple of subsets @xmath340 and a chain of subsets @xmath341 of @xmath5 $ ] , with @xmath342 . in turn
, this leads to an expression for the mbius function of @xmath1 , @xmath343 .
+ these expressions can be regarded as partial success with the computation of the zeta polynomial and the mbius function .
it would be interesting to elucidate further the question of these invariants for @xmath0 and @xmath1 , and to describe the order complexes of these posets .
4 . _ other posets of combinatorial objects with similar properties . _
+ the behavior of noncrossing partitions and restricted permutations suggests the following question : what other combinatorial objects admit a natural partial order which is self - dual and possibly , has other nice properties ?
a natural candidate is the class of _ two - stack sortable permutations _
it is known @xcite that there are as many of them with @xmath21 descents as with @xmath344 descents .
however , the poset obtained by the descent ordering is not self - dual , even for @xmath345 , so another ordering is needed .
+ similarly , the type - d noncrossing partitions and the interpolating bd - noncrossing partitions do not , in general , form self - dual posets when ordered by refinement ( see @xcite ) .
however , it may be interesting to find corresponding classes of pattern - avoiding elements in the weyl group for type d , along with an order - preserving embedding @xmath346 analogous to the type - a and b cases .
s. dulucq , s. gire , and j. west , permutations with forbidden subsequences and nonseparable planar maps , proc .
formal power series and algebr .
( florence , 1993 ) , discrete math . * 153 * ( 1996 ) 85 - 103 .
r. stanley , log - concave and unimodal sequences in algebra , combinatorics , and geometry .
graph theory and its applications : east and west ( jinan , 1986 ) , 500535 , ann .
new york acad .
, 576 , new york acad .
, new york , 1989 . | we introduce two partially ordered sets , @xmath0 and @xmath1 , of the same cardinalities as the type - a and type - b noncrossing partition lattices .
the ground sets of @xmath0 and @xmath1 are subsets of the symmetric and the hyperoctahedral groups , consisting of permutations which avoid certain patterns .
the order relation is given by ( strict ) containment of the descent sets . in each case , by means of an explicit order - preserving bijection , we show that the poset of restricted permutations is an extension of the refinement order on noncrossing partitions .
several structural properties of these permutation posets follow , including self - duality and the strong sperner property .
we also discuss posets @xmath2 and @xmath3 similarly associated with noncrossing partitions , defined by means of the excedence sets of suitable pattern - avoiding subsets of the symmetric and hyperoctahedral groups .
restricted permutations , noncrossing partitions , descents , excedences |
in a recent paper we discussed the application of the variational method to a gaussian well.@xcite we showed that harmonic variational functions yield reasonably accurate results for all well depths despite the fact that the shallow - well expansion for the ground - state variational energy does not agree with the exact expansion beyond the leading term .
students at introductory courses on quantum mechanics are commonly encouraged to solve the schrdinger equation for a particle in a one - dimensional square box with finite walls .
they thus learn that the exact solution decays exponentially as @xmath0 .
after such an experience they would find it unreasonable the choice of harmonic variational functions that do not exhibit the behavior at infinity expected for the exact solutions to the schrdinger equation for the gaussian well .
for this reason we think that it is interesting to show an alternative variational calculation based on trial functions with the expected exponential behavior at infinity . in sec .
[ sec : variational ] we develop the variational method for a general single - well potential - energy function . in sec .
[ sec : gaussian_well ] we discuss the application of the harmonic and exponential trial functions to the gaussian well , compare the approximate and exact energies for the ground and first - excited state as well as their shallow - well expansions . in sec .
[ sec : deep_well ] we compare the variational and exact deep - well expansions that we did not consider in our previous paper.@xcite finally , in sec . [ sec : conclusions ] we discuss the main results and draw conclusions .
we are interested in the solutions to the schrdinger equation @xmath1 for the dimensionless hamiltonian operator @xmath2 in our previous paper we showed how to convert the schrdinger equation into a dimensionless eigenvalue equation and we do not repeat it here.@xcite for concreteness we choose the dimensionless potential - energy function to be of the form @xmath3 , where @xmath4 and the function @xmath5 exhibits a maximum at @xmath6 and tends to zero when @xmath7 . since we restrict ourselves to single wells
we also assume that @xmath8 if @xmath9 and @xmath10 if @xmath11 although we do not make use of this condition explicitly .
we will discuss a particular example later on .
the reader will find some references about quantum - mechanical wells in our earlier paper.@xcite it is our purpose to obtain simple solutions for some states of the quantum well by means of the variational method . to this end
we choose a variational function @xmath12 that depends on a variational parameter @xmath13 .
if we assume that @xmath14 then the variational principle gives us an upper bound to the smallest eigenvalue @xmath15 and we realize that the optimum value of @xmath13 is given by the minimum of @xmath16 . if we write @xmath17 where @xmath18 and @xmath19 then @xmath20 becomes @xmath21 commonly , solving this equation for @xmath13 may not be possible or may lead to cumbersome expressions . in that case
we can resort to a simple parametric expression for the approximate energy in the form @xmath22 if @xmath23 then the states have definite parity @xmath24 and the variational principle applies to @xmath25 or @xmath26 if the variational function is even or odd , respectively .
therefore , the variational equations ( [ eq : w(a),v0(a ) ] ) apply to both the ground and first excited states .
in order to illustrate the performance of the variational method on quantum - mechanical wells we choose a particularly simple example that allows us to calculate the functions @xmath27 and @xmath28 analytically . as in our earlier paper
we select the gaussian well given by @xmath29 and begin present discussion with the calculation based on harmonic - oscillator - like variational functions.@xcite the simplest variational function for the ground state is @xmath30 and we easily obtain @xmath31 we showed that @xmath32 as @xmath33 according to@xcite @xmath34 while the exact expansion is@xcite @xmath35 note that the cubic term is missing in the approximate expansion ( [ eq : w0h_weak ] ) . we have also shown that @xmath36 is a suitable variational function for the first excited state , and in this case we have @xmath37 it is well known that @xmath38 as @xmath39 approaches a critical well strength @xmath40 from above@xcite .
we can estimate the critical strength by means of equations ( [ eq : w1h(a ) ] ) in a quite simple way .
we first obtain a positive root of @xmath41 and then @xmath42 ; the result is @xmath43 and @xmath44 . since the variational energy is an upper bound to the exact one for all @xmath39 then it is not surprising that @xmath45 .
the error for the critical strength obtained with this variational function is rather large : @xmath46 .
we may try to improve those results by means of more convenient trial functions . since @xmath47
we know that the eigenfunctions behave as @xmath48 for sufficiently large @xmath49 , where @xmath50 . in order to avoid the function @xmath49 in our calculations we work on the half positive line @xmath11 and
take into account the boundary conditions at origin : @xmath51 , @xmath52 for the even states and @xmath53 , @xmath54 for the odd ones .
thus , the simplest trial exponential function for the ground state appears to be @xmath55 the calculation of the integrals is straightforward and we obtain @xmath56}-2a\left ( 2a^{4}+a^{2}-2\right ) } \nonumber \\
w_{0e}(a ) & = & \frac{a^{2}\left\ { \sqrt{\pi } e^{a^{2}}\left ( 4a^{6}+a^{2}-2\right ) \left [ \mathrm{erf}{\left ( a\right ) } -1\right ] + 2a\left ( 2a^{4}-a^{2}+2\right ) \right\ } } { 10\left\ { \sqrt{\pi } e^{a^{2}}\left ( 4a^{6}+4a^{4}-5a^{2}+2\right ) \left [ \mathrm{erf}{\left ( a\right ) } -1\right ] + 2a\left ( 2a^{4}+a^{2}-2\right ) \right\ } } \label{eq : w_1e}\end{aligned}\ ] ] where @xmath57 is the error function . in order to obtain the @xmath39-series for @xmath58 we
first expand @xmath59 into an @xmath13-series : @xmath60 that we invert to obtain the @xmath39-series for @xmath13 @xmath61 then , we expand @xmath58 in a taylor series about @xmath62 @xmath63 and substitute the series ( [ eq : a(v0)_series_exp ] ) to obtain the shallow - well expansion @xmath64 the calculation is straightforward but extremely tedious if carried out by hand . for this reason it is a good exercise for showing the students the advantage and power of computer algebra .
such software even offer a command for obtaining the inverted series ( [ eq : a(v0)_series_exp ] ) in one step .
for sufficiently small @xmath39 this expansion is slightly better than the one derived in our earlier paper , namely eq .
( [ eq : w0h_weak ] ) , as follows from comparing the leading terms of the variational and exact @xmath39-series : @xmath65 , @xmath66 and @xmath67 . however , the cubic term that appears in the exact expansion ( [ eq : e_weak ] ) is also missing in the variational treatment based on the exponential function ( [ eq : varphi_0e ] ) . from those results
we conclude that @xmath68 for sufficiently small @xmath39 . in other words ,
the trial function with the correct asymptotic behavior at infinity yields a more accurate variational energy for sufficiently shallow wells . fig [ fig : e0 ] shows @xmath69 , @xmath58 and @xmath25 ( obtained by numerical integration@xcite ) for some values of @xmath39 .
the variational curves are almost indistinguishable in the scale of the figure .
the numerical results show that they cross at @xmath70 , where @xmath71 , so that @xmath72 if @xmath73 and @xmath74 if @xmath75 . in other words : @xmath58 is more accurate for shallow wells as argued above and @xmath69 is more accurate for deep ones .
the former inequality is consistent with the previous comparison of the leading terms of the @xmath39-expansions . in the next section
we discuss the deep - well limit .
for the first excited state we propose the simple trial function @xmath76 and obtain @xmath77 -2a\left ( 2a^{2}+5\right ) \right ] } \nonumber \\
w_{1e}(a ) & = & \frac{a^{2}\left\ { \sqrt{\pi } e^{a^{2}}\left ( 4a^{4}+8a^{2}+1\right ) \left [ \mathrm{erf}\left ( a\right ) -1\right ] + 2a\left ( 2a^{2}+3\right ) \right\ } } { 2\left\ { \sqrt{\pi } e^{a^{2}}\left ( 4a^{4}+12a^{2}+3\right ) \left [ \mathrm{erf}\left ( a\right ) -1\right ] + 2a\left ( 2a^{2}+5\right ) \right\ } } \label{eq : w1e(a)}\end{aligned}\ ] ] the variational energy vanishes at @xmath78 that leads to the approximate critical parameter @xmath79 with an error of @xmath80 , quite smaller than that for the harmonic function ( [ eq : varphi_1h ] ) .
once again we appreciate that the exponential function leads to more accurate results close to threshold ( shallow well ) .
[ fig : e1 ] shows @xmath81 , @xmath82 and @xmath26 ( obtained by numerical integration@xcite ) for a range of @xmath39 values . in this case
the variational curves are not so close each other and the crossing takes place at @xmath83 . we realize that the exponential function is preferable for the description of shallow wells as suggested by the errors in the variational critical well strengths . on the other hand , the harmonic function leads to more accurate results for deep wells
the deep - well expansion for a general well can be easily derived by means of perturbation theory.@xcite it follows from those results that the first terms of the expansion for the gaussian well are @xmath84 where @xmath85 is the harmonic - oscillator quantum number . in order to obtain the deep - well expansion for @xmath69 first note that the expression for @xmath39 in equation ( [ eq : w0h(a ) ] ) shows that @xmath86 as @xmath87 and that the leading term is @xmath88 .
therefore , we substitute @xmath89 into the expression for @xmath39 and set the unknown coefficients @xmath90 and @xmath91 in order to remove the leading powers of @xmath39 . in this way
we obtain @xmath92 finally , we substitute this equation into the expression for @xmath93 , expand and keep the leading terms ; the result is @xmath94 note that this expression agrees with ( [ eq : exact_deep_exp ] ) for @xmath95 .
proceeding exactly in the same way with equation ( [ eq : w1h(a ) ] ) we obtain @xmath96 and @xmath97 that agrees with ( [ eq : exact_deep_exp ] ) when @xmath98 .
it is clear that the harmonic variational functions are suitable for the description of deep wells which is in agreement with the results in figures [ fig : e0 ] and [ fig : e1 ] .
we have calculated the first two energy levels of the gaussian well by means of two types of variational functions . the harmonic trial functions lead to simpler expressions and their results are reasonable for all values of the well depth .
the approximate energies exhibit a slightly accurate expansion for shallow wells and a more satisfactory one for deep wells . in the latter case
they provide the first three dominant terms of perturbation theory . on the other hand ,
the expressions for the energy obtained from exponential functions are rather complicated and their results are less accurate for deep wells .
however , they are useful if one is interested in the behavior of the system near threshold ( shallow wells ) .
present results suggest that the appropriate behavior of the wavefunction at infinity does not always guarantee the greatest accuracy . in the case of a deep well
the wavefunction is strongly localized and compressed about @xmath6 and the description of the neighborhood of the origin is more important than what happens at larger @xmath49 .
a most intriguing result is that both variational approaches fail to yield the cubic term in the shallow - well expansion .
we have carried out a similar calculation for the particle in the square box with finite walls and found that the expansion of the variational energy given by the trial function ( [ eq : varphi_0e ] ) does not exhibit the cubic term that appears in the exact series . in this case
one can obtain the exact @xmath39-series for @xmath25 from the transcendental equation that determines the energy levels of that model .
we do not have a satisfactory explanation for this failure of the variational method . | we discuss two simple variational approaches to quantum wells .
the trial harmonic functions analyzed in an earlier paper give reasonable results for all well depths and are particularly suitable for deep wells . on the other hand ,
the exponential functions proposed here are preferable for shallow wells .
we compare the shallow - well expansions for both kind of functions and show that they do not exhibit the cubic term appearing in the exact series .
it is also shown that the deep - well expansion for the harmonic functions agree with the first terms of perturbation theory . |
wave collapse occurs in a wide range of physical contexts , including optics , atomic and condensed - matter physics .
generally , collapse occurs if an attractive nonlinearity exceeds a critical value .
if it is triggered at time @xmath0 , at the collapse time @xmath1 the wave amplitude asymptotically diverges at some point in space .
in practice , the amplitude divergence results in dissipation of wave energy ( or particle loss ) . the unifying theoretical framework for understanding different collapse phenomena is provided by the nonlinear schrdinger equation , which has been extensively studied for various forms of nonlinearity @xcite .
this general formalism is applied to self - focusing of light @xcite , collapse of langmuir waves @xcite and bose - einstein condensates ( becs ) @xcite , and even surface water waves @xcite . in this framework ,
wave collapse is classified as either _ strong _ or _ weak _ ( see fig .
[ fig : cartoon ] ) . in a strong collapse , a finite fraction of the wave collapses towards the singularity . on the other hand , in a weak collapse @xcite the fraction of the wave that ( in absence of dissipation ) ultimately reaches the singularity vanishes .
this has the counterintuitive practical implication that making the system _ more unstable _ , by quenching the nonlinearity further beyond the critical point , can result in _ less _ dissipation @xcite . to our knowledge
, weak collapse has not been experimentally observed in any physical system .
an atomic bec with @xmath2-wave two - body interactions is modelled by the gross - pitaevskii ( gp ) equation , with a cubic nonlinearity proportional to the scattering length @xmath3 , which can be dynamically tuned via a feshbach resonance @xcite .
the bec is prone to collapse for any @xmath4 , but a kinetic - energy barrier makes it metastable up to a critical interaction strength @xcite . if the bec becomes unstable and collapses , dissipation occurs through three - body recombination that results in particle loss .
importantly , the three - dimensional gp equation is expected to provide an example of weak collapse .
previous collapse experiments with atomic becs @xcite ( see also @xcite ) were performed in the traditional setting of a harmonic trap .
the critical point @xcite and collapse times @xcite were in general agreement with theoretical expectations @xcite , but no evidence of weak collapse was observed ; the atom loss was only seen to grow with @xmath5 @xcite .
[ btp ] in this article we study bec collapse in a new experimental setting , using a @xmath6k condensate @xcite prepared in the uniform potential of an optical - box trap @xcite ; for details of our setup see appendix a. the combination of large system size ( up to @xmath7 m ) and fine tuning of the scattering length ( with a resolution of @xmath8 , where @xmath9 is the bohr radius ) gives us a very large dynamic range : we observe metastable attractive becs with up to @xmath10 atoms , and collapse times that vary between @xmath11 and @xmath12 ms .
we demonstrate the expected scaling of the critical scattering length @xmath13 with the bec atom number @xmath14 and the system size @xmath15 , and show that the collapse time can be expressed as a universal function of the dimensionless interaction strength @xmath16 .
most importantly , we observe conclusive evidence for weak collapse , namely the counterintuitive decrease of the atom loss with increasing @xmath5 , and experimentally determine the scaling laws that govern the weak - collapse atom loss .
in sections [ sec : criticalpoint]-[sec : weakcollapse ] we address in turn : ( i ) the _ critical point _ for the collapse , ( ii ) the _ collapse dynamics _ in a system that is suddenly made unstable by an interaction quench , and ( iii ) the _ aftermath _ of the collapse , which reveals its weak nature .
the starting point for our discussion is the gp equation for a homogeneous box potential , with a heuristically added three - body loss term @xcite : @xmath17 where @xmath18 is the atom mass , @xmath19 is the three - body loss coefficient , and the boundary condition is @xmath20 at the trap walls .
we use a cylindrical box trap @xcite of variable length @xmath15 and radius @xmath21 , and always set @xmath22 , so @xmath15 is the only lengthscale characterising the system size .
we may thus rewrite eq .
( [ eq : gp ] ) in a dimensionless form , defining @xmath23 and @xmath24 , with characteristic time @xmath25 : @xmath26 where @xmath27 and @xmath28 is initially normalised to unity . in our system
@xmath29m@xmath30s@xmath31 @xcite , and for all our @xmath15 and @xmath14 values @xmath32 is very small , @xmath33 .
consequently , three - body loss is negligible in our ( meta)stable condensates .
however , if the bec collapses , significant loss occurs , providing the primary experimental signature of the collapse . neglecting the atom loss in a metastable bec , based on eq .
( [ eq : rgp ] ) the critical interaction strength , @xmath34 , can depend only on the boundary conditions , _
i.e. _ the box shape . for a family of self - similar boxes ( @xmath35 )
it should be a universal constant , so @xmath36 . to experimentally study the critical point for collapse , we prepare a stable bec at @xmath37 , then over @xmath38 s ramp the scattering length to a variable @xmath4 , and wait for @xmath39 s before turning off the trap and imaging the atoms after @xmath40 ms of time - of - flight ( tof ) expansion .
we image the cloud along the axial direction of our cylindrical trap , and for tof we jump the scattering length to @xmath41 . in fig .
[ fig : ac](a ) we show how , for a given initial @xmath14 , the final atom number depends on the negative @xmath3 .
a well defined @xmath13 is signaled by a sharp drop in the atom number . as shown in fig .
[ fig : ac](b ) , the atom loss is accompanied by a qualitative change in the appearance of the cloud in tof .
[ tbp ] , after ramping to a variable negative @xmath3 , for @xmath42 m and initial @xmath43 .
the critical scattering length @xmath13 is seen in the sharp drop in @xmath44 .
( b ) tof images taken on either side of the critical point .
( c ) variation of @xmath13 with @xmath14 , for @xmath45 m . the linear fit confirms the expected scaling @xmath46 .
( d ) variation of @xmath47 with @xmath15 .
the linear fit confirms the scaling @xmath48 .
, title="fig : " ] in fig .
[ fig : ac](c ) we plot @xmath13 for @xmath49@xmath50 m and a wide range of @xmath14 values , from @xmath51 to @xmath10 .
we clearly observe the expected scaling @xmath46 ( see also appendix b ) . in fig .
[ fig : ac](d ) we plot the measured @xmath52 versus box size and confirm the scaling @xmath53 .
we find that the dimensionless critical interaction strength is @xmath54 , where the error includes the systematic uncertainties in box size and absolute atom number calibration .
to study the collapse dynamics , we perform interaction - quench experiments @xcite .
we prepare a bec just above @xmath13 and then quench the scattering length to a variable @xmath55 to initiate the collapse . after a variable hold time @xmath56 we jump the scattering length to @xmath41 , switch off the trap , and observe the cloud in tof . as shown in the left panel of fig .
[ fig : jump](a ) , for quenches close to the critical point ( small @xmath57 ) , at @xmath1 the atom number suddenly drops to a stable lower value .
we understand this as a single collapse event . on the other hand , for large quenches [ right panel of fig .
[ fig : jump](a ) ] , the atom number appears to gradually decay until it stabilises .
such behaviour , also seen in @xcite , is understood as arising from a series of multiple ( experimentally unresolved ) collapses @xcite , and we accordingly associate @xmath1 with the onset of the atom loss .
( for further evidence for the occurrence of single and multiple discrete collapse events see appendix c. ) ( left ) and @xmath58 ( right ) ; here @xmath59 m and @xmath60 , corresponding to @xmath61 .
green bands indicate @xmath1 and its uncertainty .
( b ) typical tof images at various stages after the quench . ( c )
collapse time versus @xmath3 for six data sets taken for various @xmath15 and @xmath14 ; see legend in ( d ) .
the shaded bands indicate @xmath13 values .
( d ) universal collapse dynamics .
we plot the dimensionless collapse time versus the reduced distance from the critical point , for all six data sets . ] in fig .
[ fig : jump](b ) we show typical tof images for different times after the quench . at @xmath62 , before any change in the atom number occurs , the swelling of the cloud in tof reveals the shrinking of the wavefunction in - trap .
right after @xmath1 , within the first @xmath63 ms , we observe a shell - like structure , while at longer times we observe more irregular patterns .
we see a similar shell structure in images taken along a perpendicular direction , which implies that the outgoing atom shell is spherical . based on its size in tof
, the outer shell expands at a rate of @xmath64 mm / s , which is consistent with it reflecting off the trap walls and interfering with the central part of the cloud after @xmath63 ms . in fig .
[ fig : jump](c ) we plot @xmath1 versus @xmath3 for six data sets taken with different @xmath15 and @xmath14 values .
we observe @xmath1 values that vary between 3 and @xmath65ms . in fig .
[ fig : jump](d ) we show that all the data points fall onto a single universal curve if we plot the dimensionless collapse time , @xmath66 , versus the reduced distance from the critical point , @xmath67 .
in general , @xmath1 should also depend on @xmath32 , and between our six data sets @xmath32 varies by a factor of @xmath63 .
however , losses should start affecting the collapse dynamics only once the peak density , @xmath68 , grows by a factor @xmath69 from its initial ( pre - quench ) value .
we always have @xmath70 , so most of the dynamics should be lossless , as confirmed by the universal behaviour seen in fig .
[ fig : jump](d ) .
[ btp ] m and @xmath71 , corresponding to @xmath72 .
( a ) atom loss versus time after quenches to various @xmath3 values . for small @xmath73 ( left panel ) @xmath74 decreases with increasing @xmath5 , as expected for a weak collapse , while for large @xmath75 ( right panel ) the opposite trend is seen .
( b ) summary of atom loss for all @xmath56 and @xmath76 .
the coloured points are the data shown in ( a ) , and the points clustered around @xmath77 correspond to @xmath78 .
single - collapse loss monotonically decreases with increasing @xmath5 and extrapolates to zero for @xmath79 ( dot - dashed and solid black lines , see text ) , confirming the prediction of the weak - collapse theory . for @xmath80 single collapse does not re - stabilise the system and multiple collapse occurs .
the dashed purple line shows the equilibrium stability criterion , @xmath81 ( see text ) .
, title="fig : " ] we now turn to the aftermath of the collapse . since @xmath28 is initially normalised to unity , the _ fractional _ atom loss , @xmath74 , should be some universal function of @xmath82 and @xmath32 ; here @xmath83 is the difference between the initial ( pre - collapse ) and the final ( time - dependent ) atom number .
the counterintuitive implication of the weak - collapse theory is that @xmath74 _ decreases _ if the bec is made _ more unstable _ , by quenching @xmath3 to a more negative value . in fig .
[ fig : nlost ] we focus on one data set , for fixed @xmath59 m and @xmath71 . as we illustrate in the left panel of fig .
[ fig : nlost](a ) , close to the critical point , where we observe only single - collapse events , the atom loss indeed decreases with increasing @xmath5 , indicating weak collapse . on the other hand , as shown in the right panel of fig .
[ fig : nlost](a ) , in the regime of large quenches and multiple collapse , the atom loss in the long - time limit shows the opposite trend ; only this type of behaviour was seen in harmonic - trap experiments @xcite . in fig .
[ fig : nlost](b ) we present a consistent picture of the atom - loss trends for all @xmath84 , from @xmath85 to @xmath86 . here
we plot @xmath74 versus @xmath87 , and for each @xmath3 show @xmath74 values observed for all @xmath56 ; the points clustered around @xmath88 correspond to @xmath89 . the single - collapse regime , @xmath90 ,
is clearly identified by the small spread of the non - zero @xmath91 values .
the single - collapse atom loss clearly decreases with increasing @xmath5 , and extrapolates to zero for @xmath79 .
this is the unambiguous signature of a weak collapse .
the dot - dashed black line shows a linear extrapolation , which gives @xmath92 for @xmath93 , while the ( almost indistinguishable ) solid black line shows a power - law fit , @xmath94 . for @xmath80 ,
multiple collapse occurs , because the diminishing single - collapse atom loss does not re - stabilise the system .
however , we see that even in this regime the minimal loss we observe at each @xmath3 still follows the weak - collapse trend ( solid black line ) .
it is also instructive to plot the function @xmath81 ( dashed purple line ) ; this is atom loss such that , after a quench to a given @xmath55 , the atom number drops to the new critical value @xmath95 [ see eq .
( [ eq : dimcoef ] ) ] .
this equilibrium stability criterion is not obviously applicable in the non - equilibrium situation after the first collapse @xcite .
still , it provides a good estimate of both the point , @xmath96 , beyond which the single - collapse loss is insufficient to re - stabilise the system ( see also appendix c ) , and the long - time loss at large @xmath97 .
we now extend the study of the weak - collapse atom loss to other @xmath15 and @xmath14 values ( see fig . [
fig : weakcollapse ] ) . in this analysis
we include all @xmath3 values for which only single collapse occurs , and also those where clearly resolved single and double collapses occur ( see appendix c ) .
[ tbp ] for fixed @xmath15 and @xmath14 , for our six data sets we get an average @xmath98 .
( b ) assuming @xmath99 , we find @xmath100 ( solid black line ) , with no dependence on @xmath15 .
( c ) universal behaviour of the weak - collapse atom loss .
we plot all the single - collapse data for different @xmath3 , @xmath14 and @xmath15 versus @xmath101 , which vanishes in the limit of infinitely strong attractive interactions , @xmath102 .
the linear fit gives @xmath103 .
, title="fig : " ] writing @xmath104 for each data set with fixed @xmath15 and @xmath14 , as in fig . [
fig : nlost](b ) , we always get @xmath105 consistent with unity [ see fig .
[ fig : weakcollapse](a ) ] ; averaging over all data sets gives @xmath98 .
we then assume the form @xmath106 and study the dependence of @xmath107 on @xmath15 and @xmath14 .
as shown in fig .
[ fig : weakcollapse](b ) , on a log - log plot , we find @xmath100 , with no clear dependence on @xmath15 ; the two points taken with @xmath108 m and @xmath7 m fall onto the same line as the four points taken with @xmath59 m .
we thus experimentally find that weak - collapse atom loss is described very well by @xmath109 . from eq .
( [ eq : dimcoef ] ) , this corresponds to @xmath110 , which is indeed independent of @xmath15 , and vanishes in the limit of infinitely strong attraction , @xmath102 . in fig .
[ fig : weakcollapse](c ) we plot all our single - collapse data versus @xmath101 and confirm that it falls onto a single universal curve @xcite .
these experimentally obtained scaling laws should provide useful input for further theoretical work .
in conclusion , we have performed a comprehensive study of the collapse of an attractive bec confined in the homogeneous potential of a 3d box trap .
we have fully characterised the critical point for collapse , and the collapse dynamics of an interaction - quenched bec , finding universal behaviour in agreement with the theoretical expectations .
most importantly , we have provided conclusive experimental evidence for the counterintuitive weak collapse , and have experimentally determined weak - collapse scaling laws that should provide a useful reference point for the general theories of nonlinear wave phenomena .
our work also points to many avenues for further research .
it would be very interesting to explore quenches from a large positive @xmath3 , where the bec is initially deep in the thomas - fermi regime , and in the case of a box potential the density is uniform . in this case
it is not obvious how the condensate would spontaneously ` choose ' the position at which to collapse , or whether many local collapses would occur instead of a global one .
additionally , since the fractional atom loss can not exceed 100% , the linear trend seen in fig .
[ fig : weakcollapse](c ) can not extend to the regime of strong dissipation ( large @xmath32 ) .
it would be interesting to explore that regime using a different geometry , a different feshbach resonance , or a different atomic species .
finally , a major extension would be to perform similar experiments with 2d gases , for which a strong collapse and hence fundamentally different behaviour is expected .
we thank sarah thomas , yago del valle - incln redondo and cornelius roemer for experimental assistance , and richard fletcher , raphael lopes and andreas nunnenkamp for a critical reading of the manuscript .
this work was supported by the royal society , epsrc [ grant no .
ep / n011759/1 ] , erc ( qbox ) , afosr and aro . a.l.g . and n.n .
acknowledge support from trinity college , cambridge .
our setup is the first 3d bec box experiment with tuneable interactions .
the setup for producing harmonically trapped @xmath6k condensates is similar to our previous apparatus @xcite .
the main difference is that here we employ the gray molasses technique @xcite and directly cool @xmath6k without the need for sympathetic cooling with rubidium atoms ( see also @xcite ) .
we load the laser - cooled atoms directly into a crossed optical dipole trap ( using a @xmath111-w @xmath112-nm laser ) and achieve efficient evaporative cooling using the feshbach resonance in the @xmath113 state at 402.70(3 ) g @xcite .
this results in a quasi - pure bec of @xmath114 atoms .
we then load the atoms into the cylindrical optical box formed by blue - detuned ( @xmath115 nm ) laser light , and cancel out gravity with a magnetic field gradient , as in @xcite .
the loading procedure is essentially 100% efficient and results in a quasi - pure box - trapped bec of @xmath116 atoms .
we calibrate the atom number ( to within 10@xmath117 ) using the bec transition point of a weakly - interacting gas in a box potential @xcite .
the magnetic field at which the scattering length in the @xmath118 state vanishes was independently measured in ref .
@xcite to be . for fig .
[ fig : ac](a - c ) we calculate our @xmath3 values assuming . fitting the data in fig . [
fig : ac](c ) with a free intercept gives an intercept , which is consistent with zero within the systematic @xmath119 error due to the uncertainty in @xmath120 .
we take this to be an unbiased confirmation of the zero intercept and the expected scaling @xmath46 , and use this scaling to slightly refine the value of the zero - crossing field , to @xmath121 g. the remaining uncertainty in @xmath120 corresponds to a systematic uncertainty in our @xmath3 values of @xmath122 .
[ b ! ] for @xmath45 m , initial @xmath123 , and various closely spaced @xmath3 values ; here @xmath61 . on the right we show histograms of @xmath124 values . for @xmath125
we see two clearly resolved @xmath124 branches , corresponding to one ( upper branch ) and two ( lower branch ) collapse events .
the probability of a double collapse gradually increases with @xmath5 .
( b ) fractional atom loss versus @xmath5 on a log - log plot .
the transparent black circles show the raw data .
the coloured circles and diamonds show , respectively , the average values for the single- and double - collapse events .
the colour code is the same as in ( a ) .
the error bars show the standard deviations .
the purple dashed line shows the bec stability criterion as in fig .
[ fig : nlost](b ) and the shading shows its uncertainty . the solid black line shows the ( single - event ) weak - collapse scaling , @xmath126 .
, title="fig : " ] in fig .
[ fig : bif ] we present evidence for a gradual transition between single- and double - collapse events , which strongly supports the interpretation that an increasing number of discrete collapse events occur as @xmath5 is increased .
this data was taken with @xmath45 m and initial @xmath123 . in fig .
[ fig : bif](a ) we show the evolution of @xmath124 after a quench to various @xmath127 .
a fine scan of @xmath3 resolves a striking bifurcation of the collapse outcome .
we interpret the upper and lower branch as the result of , respectively , one and two collapse events .
as @xmath5 is increased , the probability of a double collapse gradually increases .
this crossover is highlighted in the histograms shown on the right . in fig .
[ fig : bif](b ) we show the fractional atom loss versus @xmath5 on a log - log plot . in the regime where a double collapse occurs , the single - collapse branch still clearly follows the weak - collapse scaling @xmath126 .
note that in this data set the double collapse occurs slightly closer to @xmath128 than expected from the simple equilibrium stability criterion ( purple band ) .
61ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ] + 12$12
& 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty @noop _ _ ( , , ) @noop _ _ ( , , ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.13.479 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.15.1005 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.52.1222 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.79.2604 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physreva.61.043601 [ * * , ( ) ] @noop * * , ( ) \doibase http://dx.doi.org/10.1016/0167-2789(94)90167-8 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) \doibase http://dx.doi.org/10.1016/s0375-9601(02)01385-3 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1038/408692a [ * * , ( ) ] link:\doibase 10.1103/physrevlett.86.4211 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.96.170401 [ * * , ( ) ] link:\doibase 10.1103/physreva.84.033632 [ * * , ( ) ] link:\doibase 10.1103/physreva.86.063601 [ * * , ( ) ] @noop link:\doibase 10.1103/physrevlett.90.160405 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.080401 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.108.210401 [ * * , ( ) ] link:\doibase 10.1038/nature16485 [ * * , ( ) ] @noop ( ) @noop ( ) link:\doibase 10.1103/physreva.66.011602 [ * * , ( ) ] link:\doibase 10.1103/physreva.65.033624 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0375-9601(02)00246-3 [ * * , ( ) ] link:\doibase 10.1103/physreva.67.014304 [ * * , ( ) ] link:\doibase 10.1103/physreva.68.045601 [ * * , ( ) ] link:\doibase 10.1143/jpsjs.72sc.127 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.040401 [ * * , ( ) ] link:\doibase 10.1103/physreva.71.033604 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.110.200406 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.080405 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.86.1406 [ * * , ( ) ] link:\doibase 10.1103/physreva.63.043601 [ * * , ( ) ] @noop link:\doibase 10.1103/physreva.53.r3734 [ * * , ( ) ] http://stacks.iop.org/0295-5075/104/i=6/a=63002 [ * * , ( ) ] link:\doibase 10.1103/physreva.88.053407 [ * * , ( ) ] link:\doibase 10.1103/physreva.86.033421 [ * * , ( ) ] link:\doibase 10.1103/physreva.90.033405 [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physrevlett.112.040403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.101.190405 [ * * , ( ) ] | we study the collapse of an attractive atomic bose - einstein condensate prepared in the uniform potential an optical - box trap .
we characterise the critical point for collapse and the collapse dynamics , observing universal behaviour in agreement with theoretical expectations .
most importantly , we observe a clear experimental signature of the counterintuitive weak collapse , namely that making the system more unstable can result in a smaller particle loss .
we experimentally determine the scaling laws that govern the weak - collapse atom loss , providing a benchmark for the general theories of nonlinear wave phenomena . |
about 10 per cent of white dwarfs reside in close binary systems @xcite .
the formation of white - dwarf close - binary systems commences with a pair of main - sequence stars orbiting in a wide binary .
the most massive star will evolve faster becoming a giant and transferring mass to its companion . if mass transfer is sufficiently rapid , the result will be the formation of a common envelope ( ce ) made of the outer layers of the giant .
the orbital energy of the binary can then be used to eject the envelope , the result being a binary composed of a white dwarf and a main sequence companion .
if the mass ratio of the initial components was close to unity , the binary formed after ce ejection can still be wide , whereas in the case of extreme mass ratios , the two components will spiral in to eject the envelope forming a tighter binary .
this binary , called a post common envelope binary ( pceb ) , if close enough , might become a cataclysmic variable ( cv ) if mass is transferred stably from the main sequence star to the white dwarf .
if , on the other hand , the binary is too wide to become a cv , the main sequence star will evolve into a giant and the system will undergo another ce phase .
when the envelope is ejected , the resulting binary will be composed of two white dwarfs a few solar radii apart - called a double degenerate ( dd ) . for details on the evolution and formation of dds
see nelemans et al.s recent population synthesis studies and references therein .
the study of white - dwarf close - binary stars can help us understand the elusive ce phase that they must have gone through , at least once , during their evolution .
this phase is very difficult to study in any other manner as it lasts only of the order of 1 to 100 years . for recent calculations on ce
ejection efficiencies see soker & harpaz and references therein .
the subjects of this paper , white dwarf - m dwarf binaries and dds , are , as mentioned above , the progenitors of cvs , and potential progenitors of type ia supernova respectively .
a short period dd where the combined mass of the white dwarfs exceeds the chandrasekhar mass might become a type ia supernova @xcite , but no candidates have yet been found . we should mention that this idea is still quite controversial with some theorists claiming that a dd does not become a type ia supernova @xcite and others claiming that this can be the case if one includes rotation in the calculations @xcite . the eso supernova ia progenitor survey ( spy )
has as its main goal to search systematically for massive , short period dds in the galaxy to establish their direct link with type ia supernova and it is the source of more than 100 recently discovered dd systems @xcite .
efforts are also been carried out to systematically find white - m dwarf binaries by using the sloan digital sky survey ( sdss ) resulting in more than 400 possible candidates @xcite .
schreiber & gnsicke review the evolution of the known sample of white - dwarf close binaries with main sequence companions , the pcebs , and conclude that , due to selection effects , the ( until then ) known population of pcebs is dominated by young systems with hot white dwarfs that will evolve into short period cvs ( p @xmath0 3 h ) .
the pceb sample is biased toward low mass companions as well which is the reason why they evolve into short period cvs .
in contrast , the pcebs found in the sdss @xcite , published later on , show an average white dwarf temperature significantly lower , demonstrating that to sample the full parameter space better selection criteria have to be devised .
in this paper we study 6 white dwarf close - binary stars .
four of them are found to be dds , the other two are composed of a white dwarf and an m dwarf star .
the separation of both components in these binaries is of the order of a few solar radii , as determined from their short orbital periods , so they are detached systems where no mass transfer between them takes place .
we obtain their orbital solutions and compare the results obtained with predictions drawn from population synthesis studies .
the data used in this study were taken over many years ( since 1993 ) using five different telescopes and seven different setups .
table [ obs ] gives a list of the number of spectra taken for each target at each observing campaign indicating also which telescope was used in each case .
.journal of observations .
a description of the setup used with each telescope is given in the text .
n indicates the number of spectra taken with each setup .
@xmath1 indicates high resolution spectra covering only the emission core of . the brightness in b magnitudes of each target
is also given . [ cols="<,<,<,<,<",options="header " , ]
we have obtained the orbital solution for four dd systems and two white dwarf - m dwarf binaries .
we find that the white dwarf companions for the four dds studied contribute between 10 and 20 per cent of the total luminosity and remain undetected .
the masses and periods obtained for these systems agree with theoretical mass period distributions ( based on nelemans et al . ) .
in the case of the white dwarf - m dwarf binaries we have been able to measure the motion of both components .
we find that there are signatures of strong irradiation of the surface of the m dwarf component in .
the white dwarf masses calculated from their gravitational redshift are unusually high compared to those measured by fitting their hydrogen lines with stellar atmosphere models , and could be the result of the evolution of giant stars with masses between 1.25 and 3.5 that went through a ce phase as a result of tidal interaction .
these two binaries are young pcebs that will evolve into cvs , although not within a hubble time , with orbital periods below the period gap .
lmr was supported by a pparc post - doctoral grant and by nwo - vidi grant 639.042.201 to p.j .
groot during the course of this research .
gn is supported by pparc and an nwo - veni grant .
trm acknowledges the support of a pparc senior research fellowship .
rn acknowledges support by a pparc advanced fellowship . the authors would like to thank m. r. schreiber for his help with implementing the calculations for this paper .
the reduction and analysis of the data were carried out on the southampton node of the starlink network .
we thank patt for their support of this program . | we determine the orbits of four double degenerate systems ( dds ) , composed of two white dwarfs , and of two white dwarf
m dwarf binaries .
the four dds , , , , and , show orbital periods of 1.157155(5 ) d , 1.15674(2 ) d , 6.26602(6 ) d and 5.0846(3 ) d respectively .
these periods combined with estimates for the masses of the brighter component , based on their effective temperatures , allow us to constrain the masses of the unseen companions .
we estimate that the upper limit for the contribution of the unseen companions to the total luminosity in the four dds ranges between 10 and 20 per cent . in the case of the two white dwarf - m dwarf binaries , and
, we calculate the orbital parameters by fitting simultaneously the absorption line from the white dwarf and the emission core from the m - dwarf .
their orbital periods are 0.337083(1 ) d and 0.741226(2 ) d respectively .
we find signatures of irradiation on the inner face of s companion .
we calculate the masses of both components from the gravitational redshift and the mass - radius relationship for white dwarfs and find masses of 0.75 0.78 and 0.61 0.64 for and respectively .
this indicates that the stars probably reached the asymptotic giant branch in their evolution before entering a common envelope phase .
these two white dwarf - m dwarf binaries will become cataclysmic variables , although not within a hubble time , with orbital periods below the period gap .
latexl-.36em.3ex-.15em t-.1667em.7ex-.125emx = [ firstpage ] binaries : close binaries : spectroscopic white dwarfs |
Subsets and Splits